2011-10-25T00:01:52 *** vladimirfol has quit IRC (Ping timeout: 265 seconds) 2011-10-25T00:06:46 so has anyone started reading about the multi-agent online traveling salesperson problem? 2011-10-25T00:06:47 *** antimatroid has joined #aichallenge 2011-10-25T00:07:02 *** antimatroid1 has quit IRC (Read error: No route to host) 2011-10-25T00:08:55 *** skalas has joined #aichallenge 2011-10-25T00:09:01 *** dauryg has quit IRC (Quit: Page closed) 2011-10-25T00:11:37 *** bmh_ has joined #aichallenge 2011-10-25T00:11:37 *** bmh_ has joined #aichallenge 2011-10-25T00:13:33 *** antimatroid has quit IRC (Ping timeout: 260 seconds) 2011-10-25T00:14:39 *** bmh has quit IRC (Ping timeout: 248 seconds) 2011-10-25T00:14:39 *** bmh_ is now known as bmh 2011-10-25T00:16:56 oh, when you meet enemy ants, you DO get to know which team they're on?(eg. you can tell the difference between 3vs5 and 3vs2vs2vs1? 2011-10-25T00:18:27 *** guest_____ has quit IRC (Quit: Page closed) 2011-10-25T00:20:01 *** antimatroid has joined #aichallenge 2011-10-25T00:20:39 *** otzi has quit IRC (Ping timeout: 255 seconds) 2011-10-25T00:20:46 anti, are you one of the organizers of this? and/or are you spending a lot of time, and helping noobish people? 2011-10-25T00:23:19 *** dvladim has quit IRC (Ping timeout: 258 seconds) 2011-10-25T00:23:47 when my bot knows it's won.. I'm gonna have it spell out smack talk 2011-10-25T00:24:01 Is my A* slow? For tutorial1.map finding the path from start (28,19) to (14,19) takes me ~215ms 2011-10-25T00:24:16 whoa. yes. 2011-10-25T00:24:19 what language? 2011-10-25T00:24:22 python 2011-10-25T00:24:35 *** TheDigitalNinja has quit IRC (Ping timeout: 265 seconds) 2011-10-25T00:24:38 what's a good target? 2011-10-25T00:24:50 i mean, good time for that path? 2011-10-25T00:24:55 not sure, haven't done it in python 2011-10-25T00:25:05 there's no way it should take that long 2011-10-25T00:25:31 yeah.. is it exploring everything? what's your heuristic, and what are you using for a priority queue? 2011-10-25T00:25:42 can you printout the number of nodes explored? 2011-10-25T00:26:13 it's a grid so you can use manhattan distances 2011-10-25T00:26:14 manhattan distance. I'm using aima-python's priority queue 2011-10-25T00:26:29 *** Areks|2 has quit IRC (Ping timeout: 252 seconds) 2011-10-25T00:26:35 aima as in norvig's book? 2011-10-25T00:26:56 yes. http://aima-python.googlecode.com/svn/trunk/utils.py, see class PriorityQueue 2011-10-25T00:28:05 something's wrong with the algorithm then. count up the number of nodes explored. 2011-10-25T00:28:54 it should explore something like 19 or 20 2011-10-25T00:28:59 161 nodes popped 2011-10-25T00:29:09 Is the order in which we will receive "w" "h" "a" and "d" messages always the same? 2011-10-25T00:29:35 kfjlasd: print out the coords each time? see if you have dupes? 2011-10-25T00:29:48 also, 215ms for 161 nodes? whaat 2011-10-25T00:30:09 are you using some fancy heuristic? 2011-10-25T00:30:19 < kfjlasd> manhattan distance. 2011-10-25T00:30:33 *** kegi has joined #aichallenge 2011-10-25T00:32:04 I'm using manhattan distance: h(n) = D * (abs(n.x-goal.x) + abs(n.y-goal.y)), with D=1. from http://theory.stanford.edu/~amitp/GameProgramming/Heuristics.html#manhattan-distance 2011-10-25T00:32:23 you'll need to check for wrapping there 2011-10-25T00:32:39 but that's not the issue at hand here 2011-10-25T00:32:45 kfjlasd: for the love of god please indes using [row][col], [x][y] is horribly confusing 2011-10-25T00:32:55 ya, my bad, can't read 2011-10-25T00:33:01 like a1k0n said, try printing out the nodes 2011-10-25T00:33:43 *** Stasmo has joined #aichallenge 2011-10-25T00:33:50 hey friend pals 2011-10-25T00:33:51 foob: no, the order can change 2011-10-25T00:33:57 if that doesn't reveal anything, try putting in timer log, to see what's taking up the time 2011-10-25T00:34:00 *** amstan has quit IRC (Ping timeout: 240 seconds) 2011-10-25T00:34:08 My nodes: http://pastebin.com/nc4eeszv 2011-10-25T00:35:43 *** twymer has quit IRC (Ping timeout: 256 seconds) 2011-10-25T00:36:36 looks like lot of duplicates, might be a bug in tracking/checking for explored node 2011-10-25T00:36:36 well there's something wrong with the pq or the distance in it 2011-10-25T00:36:48 janzert: thanks. 2011-10-25T00:36:56 *** delt0r_ has quit IRC (Ping timeout: 258 seconds) 2011-10-25T00:37:17 not only are there dups, but why on earth did it go from (23,20) to (28,21) 2011-10-25T00:37:39 and are you maintaining a closed set? 2011-10-25T00:38:06 what do you mean exactly? 2011-10-25T00:38:19 don't add a location to the queue more than once 2011-10-25T00:38:20 in A*, usually there's an open set and a closed set, the latter being the set of nodes already explored 2011-10-25T00:38:25 http://en.wikipedia.org/wiki/A*_search_algorithm#Pseudocode 2011-10-25T00:38:34 so each time you try an add, you check the closed set first 2011-10-25T00:38:43 yes, i have a set of explored states 2011-10-25T00:39:08 don't think you need open and closed sets on a grid? 2011-10-25T00:39:18 sure you do 2011-10-25T00:39:27 you don't need closed sets in continuous space usually 2011-10-25T00:39:35 nah, cause the first time you come across a location it's going to be the fastest you can get to it 2011-10-25T00:39:43 so why bother checking if you find a shorter path later? 2011-10-25T00:39:51 just mark a location as added when you add it to the queue and don't add it again 2011-10-25T00:40:11 only works as long as you don't pop the queue 2011-10-25T00:40:13 right, you can impement the "closed set" by just putting some info into the square on the grid 2011-10-25T00:40:45 so you only really need closed and not open? 2011-10-25T00:41:01 you need open -- that's your queue of stuff to check 2011-10-25T00:41:02 open is for when added to the queue but not taken off yeah? 2011-10-25T00:41:07 closed just keeps you from looping 2011-10-25T00:41:18 *** skalas has quit IRC (Quit: Leaving) 2011-10-25T00:41:18 in case two equivalent paths go to the same place 2011-10-25T00:41:23 yeah okay 2011-10-25T00:44:40 *** twymer has joined #aichallenge 2011-10-25T00:46:39 *** analyst74 has quit IRC (Ping timeout: 244 seconds) 2011-10-25T00:49:14 *** delt0r_ has joined #aichallenge 2011-10-25T00:53:15 *** twymer has quit IRC (Ping timeout: 252 seconds) 2011-10-25T00:53:33 *** analyst74 has joined #aichallenge 2011-10-25T00:55:38 *** cyphase has quit IRC (Read error: Operation timed out) 2011-10-25T01:02:47 *** kfjlasd has quit IRC (Ping timeout: 265 seconds) 2011-10-25T01:07:49 *** Garf has joined #aichallenge 2011-10-25T01:08:04 is there a way to use boost on the server? apparently it is installed according to some threads in the forums 2011-10-25T01:09:52 hmm, they are actually really old threads from previous contests 2011-10-25T01:10:00 *** dgroft has joined #aichallenge 2011-10-25T01:10:44 note that if you don't need a boost .lib, you can simply ship the code along 2011-10-25T01:10:52 whats spawnradius2? 2011-10-25T01:11:11 I know, but most of the stuff doesn't fit in the 2MB 2011-10-25T01:11:33 *** pvarga_ has joined #aichallenge 2011-10-25T01:11:49 lost password. any managers avail to reset password for me? 2011-10-25T01:12:24 *** pvarga has quit IRC (Ping timeout: 240 seconds) 2011-10-25T01:12:24 *** pvarga_ is now known as pvarga 2011-10-25T01:16:46 jesus christ 2011-10-25T01:16:54 i just spend like half a day trying to solve this stupid little thing 2011-10-25T01:16:58 and it was like 2 lines of easy 2011-10-25T01:17:05 *** Puj has quit IRC (Ping timeout: 276 seconds) 2011-10-25T01:18:04 alright, I made a forum post about it so we'll see what happens :) 2011-10-25T01:20:44 spawnradius is the "food eating radius" apparently 2011-10-25T01:22:38 *** kegi has quit IRC (Quit: Page closed) 2011-10-25T01:26:06 for the record ants used to spawn where the food was 2011-10-25T01:26:09 so that's why it's called that 2011-10-25T01:26:22 it used to make more sense 2011-10-25T01:27:06 * janzert just expanded the description on the specification page 2011-10-25T01:28:14 *** ants15 has joined #aichallenge 2011-10-25T01:28:58 *** amstan has joined #aichallenge 2011-10-25T01:28:58 *** ChanServ sets mode: +o amstan 2011-10-25T01:29:11 contestbot: messages? 2011-10-25T01:29:12 amstan: I'm sorry Dave, err amstan; I cannot 'messages?'. 2011-10-25T01:29:59 *** irchs has quit IRC (Quit: irchs) 2011-10-25T01:32:04 Am I still popping too many nodes? http://pastebin.com/mzPPntBv 2011-10-25T01:36:28 *** ants15 has quit IRC (Quit: Page closed) 2011-10-25T01:37:31 @later tell McLeopold Can you look at issue #326 when you get a chance? I think it should be an easy one for you to resolve :) 2011-10-25T01:37:31 janzert: As you wish. 2011-10-25T01:37:32 Unknown command, try @list 2011-10-25T01:37:55 uh oh, we got two @ listeners now? 2011-10-25T01:38:41 > error "silly bot" 2011-10-25T01:38:51 >error 2011-10-25T01:38:56 :\ 2011-10-25T01:39:00 *** Chris_0076 has quit IRC (Remote host closed the connection) 2011-10-25T01:39:14 ok, who brought lambdabot? 2011-10-25T01:39:44 *** Chris_0076 has joined #aichallenge 2011-10-25T01:39:49 * janzert eyes lambdabot suspiciously 2011-10-25T01:39:56 it's a spy 2011-10-25T01:40:05 @aarossig: test 2011-10-25T01:40:05 Unknown command, try @list 2011-10-25T01:40:05 amstan: An error occured while trying to show the previous error. 2011-10-25T01:40:14 it's this: http://www.haskell.org/haskellwiki/Lambdabot 2011-10-25T01:40:14 yep, they both complain 2011-10-25T01:40:35 ooh, a haskel spy no less 2011-10-25T01:40:36 thestinger: i know what it is, the problem is that both of them use the same character as a command 2011-10-25T01:40:59 jmcarthur: any idea who brought lambdabot? 2011-10-25T01:41:30 haha. 2011-10-25T01:41:50 @list 2011-10-25T01:41:50 a1k0n: AIChallenge, Admin, Alias, Anonymous, BadWords, Channel, ChannelLogger, ChannelStats, Config, Ctcp, Dict, Dunno, Factoids, Filter, Format, Games, Google, Herald, Internet, Later, Limiter, Math, Misc, Network, News, Note, Owner, Plugin, Quote, QuoteGrabs, RSS, Reply, Scheduler, Seen, Services, ShrinkUrl, Status, String, Success, Time, Todo, Topic, URL, Unix, User, Utilities, and Web 2011-10-25T01:41:50 http://code.haskell.org/lambdabot/COMMANDS 2011-10-25T01:42:18 *** r-bit has joined #aichallenge 2011-10-25T01:42:57 *** lambdabot has left #aichallenge 2011-10-25T01:44:06 ok, he left 2011-10-25T01:44:47 the best part is 2011-10-25T01:44:52 i dont know anything about writing AI 2011-10-25T01:44:58 *** X3 has joined #aichallenge 2011-10-25T01:45:11 Stasmo: think of it like solving a problem 2011-10-25T01:46:38 *** Areks has joined #aichallenge 2011-10-25T01:47:22 right now i am thinking, should i make an AI that tells my ants things and lets them decide what to do, or should i make the decisions for them? 2011-10-25T01:47:46 *** randll has joined #aichallenge 2011-10-25T01:47:57 Stasmo: you have to tell each ant where to go 2011-10-25T01:48:18 i tell the ants where they can go, and how good for the team it would be for them to go there 2011-10-25T01:48:21 *** foob has quit IRC (Quit: brain explosions) 2011-10-25T01:48:28 and they decide for themselves based on what i tell them 2011-10-25T01:48:49 the ants don't decide for themselves, your program decides for all of them at once 2011-10-25T01:48:57 i also tell them things like how close their buddies are and how close the enemies are 2011-10-25T01:49:13 i have given a bit of the ai to the ants 2011-10-25T01:49:22 * amstan is confused 2011-10-25T01:49:45 well, i gave the ants an array that holds orders and their priorities 2011-10-25T01:50:25 and basically i just spend each update telling the ants what to do, then cycling through each ant and getting it to figure out what the highest priority order is 2011-10-25T01:51:44 *** Palmik has joined #aichallenge 2011-10-25T01:51:50 priorities are based off of the decided status of the game 2011-10-25T01:51:51 Stasmo: sounds like a good idea 2011-10-25T01:52:08 if i am trying to grow an army, higher priority is placed on the gathering of foooood 2011-10-25T01:52:26 Is the priority different for every ant? 2011-10-25T01:52:44 if i want to crush bases, i can change the status to kill mode or something and then higher priority goes to finding and killing enemies 2011-10-25T01:52:46 *** grom358 has quit IRC (Quit: Ex-Chat) 2011-10-25T01:52:46 or do they all pick from the same array 2011-10-25T01:52:58 they all pick from the same array 2011-10-25T01:53:02 currently 2011-10-25T01:53:31 The currently it sounds like a good idea to use a priority queue for that 2011-10-25T01:53:33 although getting them to decide their own priorities would be a good idea 2011-10-25T01:54:32 it always has the best order right in front of the list 2011-10-25T01:55:41 *** wiz3kid has joined #aichallenge 2011-10-25T01:55:52 hi 2011-10-25T01:57:15 *** wiz3kid has quit IRC (Client Quit) 2011-10-25T01:58:01 currently i just shove em all in there, but a priority queue for each ant is a good idea 2011-10-25T01:58:05 thank you 2011-10-25T01:58:53 now we can have nyan cats flying around while watching ant games: http://nyanit.com/aichallenge.org 2011-10-25T01:59:50 *** pvinis has joined #aichallenge 2011-10-25T01:59:57 hello 2011-10-25T02:00:20 are we supposed to change only the MyBot file? 2011-10-25T02:00:28 or the ants file as well? 2011-10-25T02:00:31 pvinis: nope, change whatever you want 2011-10-25T02:00:51 oh ok 2011-10-25T02:01:22 because there are some functions in ants.py that are not in the ants.rb file 2011-10-25T02:01:56 pvinis: if you can fix that, we could update the starter packages from the main site to include your fix 2011-10-25T02:02:03 if you look through all the code, you can see that the way they got this to work in several languages is because these different technologies interact by passing eachother strings 2011-10-25T02:02:15 i will later today 2011-10-25T02:02:19 and then parsing passed strings 2011-10-25T02:02:38 so make sure you pass the right string when given the right info and you can do whatever you want with this code 2011-10-25T02:02:41 yes i saw that, but some useful things are missing from the ruby files 2011-10-25T02:02:45 it's brilliant! 2011-10-25T02:03:09 i'm doing mine in php 2011-10-25T02:03:17 nice 2011-10-25T02:03:30 i started today. looks fun :p 2011-10-25T02:03:34 its my first time 2011-10-25T02:03:39 *** bobbydroptable has joined #aichallenge 2011-10-25T02:05:27 programming is really fun if you really get into it 2011-10-25T02:05:33 i've noticed 2011-10-25T02:05:50 so what's our memory limit? 2011-10-25T02:06:27 i've heard it is 1gb 2011-10-25T02:06:35 also, what sort of processor time will get? 2011-10-25T02:07:10 say, in multiples of the processing power i get with my laptop? 2011-10-25T02:07:26 *** r-bit has quit IRC (Quit: Page closed) 2011-10-25T02:07:27 1 2011-10-25T02:07:27 probably not 2011-10-25T02:07:33 1,.5 maybe 2011-10-25T02:09:30 okay 2011-10-25T02:09:37 so i can't go overboard with the map 2011-10-25T02:09:43 yup you can't 2011-10-25T02:09:50 math* 2011-10-25T02:09:59 don't try and A* from 1 end of the map to the other, for example 2011-10-25T02:10:09 lol 2011-10-25T02:10:15 without some macro level path 2011-10-25T02:10:21 that would be fun! a* 2011-10-25T02:10:24 but that's easy! 2011-10-25T02:10:34 from the upper left corner to the lower right? 2011-10-25T02:10:35 see, i saw that on the forum 2011-10-25T02:10:38 up, left. done. 2011-10-25T02:10:39 >.> 2011-10-25T02:10:40 lol 2011-10-25T02:10:47 okay "over large distances" then 2011-10-25T02:11:22 ants dont know anything about Dijkstra 2011-10-25T02:11:25 they're only ants 2011-10-25T02:11:35 the largest manhattan distance, disregarding the presence of water, is from the center to any corner 2011-10-25T02:11:49 thanks to the maps being toroidal 2011-10-25T02:12:10 hi guys. maybe someone knew.. is there any chance that will get additional servers to run competition? "Next game should be within 177.5 minutes. " looks kinda frustrating... 2011-10-25T02:15:53 randll, use the tcp servers 2011-10-25T02:16:16 randll, http://ants.fluxid.pl/howto and http://ash.webfactional.com/howto 2011-10-25T02:17:34 not exactly what i wanted, but thanks :) 2011-10-25T02:20:47 *** he_the_great has quit IRC (Ping timeout: 276 seconds) 2011-10-25T02:21:49 *** deltree_ has joined #aichallenge 2011-10-25T02:22:06 aichallenge: janzert epsilon * ra7c1982 / website/index.php : Expand front page game display - http://git.io/eTviVQ 2011-10-25T02:22:45 *** g0llum has joined #aichallenge 2011-10-25T02:22:57 randll, given that the games can take 1000 turns and have 10 players, each of which can use 0.5 seconds of processing power, which comes out to something like 84 cpu minutes, i think it's pretty good that we have 2 games / minute :P 2011-10-25T02:23:05 *** lundgren has joined #aichallenge 2011-10-25T02:24:25 *** ltriant has quit IRC (Quit: Computer has gone to sleep) 2011-10-25T02:24:27 actually, it's 4 games / minute atm 2011-10-25T02:26:04 amstan: it probably doesn't matter too much, but adding aichallenge.org site specific stuff to the github repo could cause other people trying to set it up themselves problems 2011-10-25T02:26:05 basically, if you want to see your bot play a lot of games against other players fast, you'll want to use the tcp servers 2011-10-25T02:26:53 janzert: hmm? 2011-10-25T02:27:02 can't...stay...awake, help! 2011-10-25T02:27:08 *slaps with a trout* 2011-10-25T02:27:22 *trout falls asleep* 2011-10-25T02:27:24 just noticed the forum redirect that was added to the apache config template 2011-10-25T02:27:47 gotta stay up for 2 more hours and then check on my car at 4:30am 2011-10-25T02:27:57 *** formics has joined #aichallenge 2011-10-25T02:27:58 janzert: oh, ok 2011-10-25T02:28:46 is it possible to add new organizations to the affiliation list? 2011-10-25T02:29:09 *** the-mgt has quit IRC (Quit: Pogodan - http://pogodan.com) 2011-10-25T02:32:07 *** djr_ has quit IRC (Read error: Connection reset by peer) 2011-10-25T02:33:56 what's a good way to do logging in PHP? 2011-10-25T02:35:50 *** the-mgt has joined #aichallenge 2011-10-25T02:36:53 *** GeorgeSebastian has joined #aichallenge 2011-10-25T02:37:35 *** GeorgeSebastian has quit IRC (Max SendQ exceeded) 2011-10-25T02:38:22 *** cybsy has joined #aichallenge 2011-10-25T02:38:28 *** GeorgeSebastian has joined #aichallenge 2011-10-25T02:38:46 Stasmo: write to stderr and use -E as an argument to playgame.py 2011-10-25T02:39:06 thank you 2011-10-25T02:40:00 you can set the log dir with '--log_dir whatever' and with -E it will log all the stderr output you do 2011-10-25T02:41:49 *** GeorgeSebastian has quit IRC (Client Quit) 2011-10-25T02:42:41 *** Zacru has joined #aichallenge 2011-10-25T02:42:44 *** GeorgeSebastian has joined #aichallenge 2011-10-25T02:44:31 yay, I finally got A* search working in C++ 2011-10-25T02:44:32 :D 2011-10-25T02:44:50 in a toy program though... I just need to rewrite the whole thing to work with my bot now :P 2011-10-25T02:46:27 all that emphasis on pathing 2011-10-25T02:46:44 what about the fighting? 2011-10-25T02:46:53 and the actual decision making? 2011-10-25T02:47:02 i am worried about my ants 2011-10-25T02:47:34 *** ikaros has joined #aichallenge 2011-10-25T02:47:35 one thing at a time :) 2011-10-25T02:48:37 I did some research and it looks like game trees are a neat way to handle tactical fighting stuff, I've never done stuff like that before though 2011-10-25T02:48:58 *** aerique has joined #aichallenge 2011-10-25T02:49:28 *** onensora has joined #aichallenge 2011-10-25T02:50:27 *** notosonerdysunny has joined #aichallenge 2011-10-25T02:50:59 Hi Everybody,Bad submission key found 2011-10-25T02:51:24 I keep getting "Bad submission key found " when I submit even of the starter packages that I downloaded .. Can somebody help... 2011-10-25T02:51:36 I keep getting "Bad submission key found " when I submit even one of the starter packages that I downloaded .. Can somebody help... 2011-10-25T02:52:03 notosonerdysunny: Random guess, maybe your session timed out? and you need to log in again? 2011-10-25T02:52:37 yeah, try refreshing the submit page then uploading 2011-10-25T02:53:10 *** crem has left #aichallenge 2011-10-25T02:53:18 *** kaemo has quit IRC (Ping timeout: 245 seconds) 2011-10-25T02:53:27 *** porco has joined #aichallenge 2011-10-25T02:53:34 *** formics has quit IRC (Remote host closed the connection) 2011-10-25T02:56:05 *** cybsy has quit IRC (Ping timeout: 252 seconds) 2011-10-25T02:56:06 *** Puj has joined #aichallenge 2011-10-25T02:57:40 How do you print to stderr in Python? 2011-10-25T02:58:30 assuming python 3, print("stuff", file=sys.stderr) 2011-10-25T02:58:42 ty 2011-10-25T02:58:46 *** Stasmo has quit IRC (Ping timeout: 265 seconds) 2011-10-25T02:58:54 in python 2 you can just 'from __future__ import print_function', cause it's nicer 2011-10-25T02:58:59 python2 print >>sys.stderr, "blah" 2011-10-25T03:01:02 So could I just write something like def printE(message): 2011-10-25T03:01:27 print(message, file=sys.stderr) 2011-10-25T03:01:39 ? 2011-10-25T03:01:48 Nope, that's forbidden by guido. 2011-10-25T03:01:59 :) 2011-10-25T03:02:04 haha 2011-10-25T03:02:21 Why do you say that? 2011-10-25T03:02:46 Zepp: Just joking. 2011-10-25T03:03:02 *** GeorgeSebastian has quit IRC (Ping timeout: 276 seconds) 2011-10-25T03:06:10 So how do you limit the depth of an A* search? Just find the distance to the starting node and stop if it's > than some max range? 2011-10-25T03:07:04 dangit, pathfinding is so tricky 2011-10-25T03:07:54 I've got breadth first searches working, except that they're not always accurate, and they're extremely resource intensive 2011-10-25T03:08:29 *** porco` has joined #aichallenge 2011-10-25T03:08:31 @deltree_ I use BFS too, just never search for things that are too far 2011-10-25T03:08:32 analyst74: User error -- Replace user. 2011-10-25T03:08:54 analyst74: it doesn't matter it seems, I limited it to 8 moves 2011-10-25T03:09:23 8 is huge, if you do that on a non-trivial number of ants 2011-10-25T03:09:35 I timed out when I hit 4 2011-10-25T03:09:35 *** porco has quit IRC (Ping timeout: 248 seconds) 2011-10-25T03:09:44 deltree_: what language? 2011-10-25T03:09:48 c# 2011-10-25T03:10:05 now, I'm still working with the starter package code, so it's pretty rough code all in all 2011-10-25T03:10:13 I'm just trying to get the thing "working" 2011-10-25T03:10:27 I guess I should ask: How much faster is c++ than python? 2011-10-25T03:11:02 quite a bit, although I would guess better strategy is more important 2011-10-25T03:11:16 'Cause I'm pretty comfortable with python, and am just starting c++. I'm not sure I 2011-10-25T03:11:20 *** xar0l_ has quit IRC (Ping timeout: 265 seconds) 2011-10-25T03:11:37 Zacru: a _lot_ faster, if you're using the same algorithms 2011-10-25T03:11:46 I'd be able to write the c++ code 2011-10-25T03:11:53 I saw how one guy solved the problem of commanding multiple ants in a recent game 2011-10-25T03:12:20 he just gave them all the same command until the task was complete, many of them got held up on walls and things, but it didn't matter 2011-10-25T03:13:10 wow i do bfs search on single targets like hills for about 30 range and don't time out 2011-10-25T03:13:21 *** cirno_the_greate has quit IRC (Remote host closed the connection) 2011-10-25T03:13:31 or i think even 50 on occassion 2011-10-25T03:13:39 I suppose I'll just have to rewrite the code for better speed 2011-10-25T03:14:00 hilll razing is special, there ain't going to be many hills 2011-10-25T03:14:06 true 2011-10-25T03:14:18 but for food i use 2*viewradius 2011-10-25T03:14:26 thats 14 or 16 or so 2011-10-25T03:14:27 plus, I don't think my bfs is even working correctly, it could be using up an unusually high amount of resources 2011-10-25T03:14:39 wow, that's pretty good ikaros 2011-10-25T03:14:54 *** Fandekasp has joined #aichallenge 2011-10-25T03:15:04 food gathering works nice so far.. but the rest of my bot is still very dumb lol 2011-10-25T03:15:09 and it's not always accurate, which sounds broken to me 2011-10-25T03:15:16 true 2011-10-25T03:15:33 is just logged the maps i calculated to a file 2011-10-25T03:15:36 to check them 2011-10-25T03:15:43 anyway, I have another hour and 15 minutes to kill, so I think I'm going to try to nap and get up in time.... 2011-10-25T03:16:18 python is about 25x-50x slower than C/C++. C# about 3x slower. 2011-10-25T03:16:32 generally speaking, on average, depends on lots of factors blabla 2011-10-25T03:18:08 cpython is 25x-50x slower* 2011-10-25T03:18:11 http://speed.pypy.org/ :P 2011-10-25T03:18:25 for this contest I guess we're stuck with cpython though 2011-10-25T03:18:29 you can't use pypy afaik 2011-10-25T03:18:37 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-10-25T03:19:20 i use lua and it's quite fast.. especially with luajit 2011-10-25T03:20:02 funny to see python in #1 for a 25-50x slower language :P 2011-10-25T03:20:37 good algos always beat brute force i guess :) 2011-10-25T03:20:38 development speed matters too. particularly as prolly noone really knows yet what good strategy is 2011-10-25T03:22:21 dear, and I can't even do the basic A* stuff xD 2011-10-25T03:23:38 http://www-cs-students.stanford.edu/~amitp/gameprog.html#Paths good resource 2011-10-25T03:24:04 krishna: I tried refreshing and uploading .. it didn't help 2011-10-25T03:24:09 Yes I've read it yesterday 2011-10-25T03:24:14 janzert: I tried refreshing and uploading .. it didn't help 2011-10-25T03:25:09 *** analyst74 has quit IRC (Ping timeout: 256 seconds) 2011-10-25T03:26:48 *** lundgren has quit IRC (Ping timeout: 265 seconds) 2011-10-25T03:26:54 *** Zacru has quit IRC (Quit: Page closed) 2011-10-25T03:27:09 either your session is ending between your getting the upload page and submitting it or somehow your browser is not sending or changing a hidden form field specific to your account that prevents some attacks 2011-10-25T03:27:39 not really sure what else to tell you to try other than maybe restarting your browser, or else maybe try a different one 2011-10-25T03:28:48 blah ai-class probability is doing my head in... 2011-10-25T03:29:14 is there some command line way of submitting .. some script or something like that.. 2011-10-25T03:29:25 janzert: ^ 2011-10-25T03:29:44 not that I know of 2011-10-25T03:30:05 someone made one for the last contest, but I haven't seen one for this yet 2011-10-25T03:30:34 oh ok .. I think that would be very convenient and usefull... 2011-10-25T03:31:02 *** Blkt has joined #aichallenge 2011-10-25T03:37:16 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-10-25T03:38:23 *** SharkMonkey has joined #aichallenge 2011-10-25T03:38:49 Does --nolaunch prevent the html file from being generated with the replay data? 2011-10-25T03:39:10 Can I have the web browser not be launched, but that file be generated? 2011-10-25T03:46:29 Got it. --nolaunch with an explicit --html= flag did the trick 2011-10-25T03:47:09 Didn't work when I told it game_logs/replay.0.html as it already uses --log_dir 2011-10-25T03:49:26 *** danielharan has quit IRC (Remote host closed the connection) 2011-10-25T03:49:50 *** schnoid has joined #aichallenge 2011-10-25T03:51:56 *** notosonerdysunny has quit IRC (Ping timeout: 265 seconds) 2011-10-25T03:53:59 hi all, I noticed in the tutorial it said there's supposed to be an isVisible method in the Ant class (in the starter package). 2011-10-25T03:54:04 am i missing something? 2011-10-25T03:55:14 Is there any reason the system only plays 2 games per minute? How many machines are working on it? 2011-10-25T03:55:52 according to this page http://aichallenge.org/server_stats.php - from 5 to 11 2011-10-25T03:56:47 randll: thanks 2011-10-25T04:00:28 *** sir_macelon has joined #aichallenge 2011-10-25T04:01:05 good morning everyone 2011-10-25T04:01:13 hihi 2011-10-25T04:02:02 McLeopold: I even wanted to write it by my own, but surely I'll look into it. 2011-10-25T04:02:50 *** Palmik has quit IRC (Ping timeout: 276 seconds) 2011-10-25T04:04:30 *** porco has joined #aichallenge 2011-10-25T04:07:03 *** kaemo has joined #aichallenge 2011-10-25T04:07:37 *** kiv has joined #aichallenge 2011-10-25T04:07:44 *** tdubellz has quit IRC (Read error: Operation timed out) 2011-10-25T04:09:07 *** tdubellz has joined #aichallenge 2011-10-25T04:09:23 *** tdubellz is now known as Guest58853 2011-10-25T04:10:47 *** Palmik has joined #aichallenge 2011-10-25T04:13:35 *** porco has quit IRC (Ping timeout: 248 seconds) 2011-10-25T04:14:28 *** Akranis has joined #aichallenge 2011-10-25T04:14:42 *** bearoff has joined #aichallenge 2011-10-25T04:15:37 *** schnoid has quit IRC (Ping timeout: 265 seconds) 2011-10-25T04:17:04 *** skunx has joined #aichallenge 2011-10-25T04:17:27 i am looking at the tutorial 2011-10-25T04:17:30 step 2 2011-10-25T04:17:49 where it says to replace the default move with a new one 2011-10-25T04:17:49 pvinis: which language? 2011-10-25T04:17:55 python 2011-10-25T04:18:10 i'm working on ruby, but looking at python tutorial 2011-10-25T04:18:18 which is the default move? 2011-10-25T04:18:37 *** hacklash has joined #aichallenge 2011-10-25T04:18:39 *** ikaros has joined #aichallenge 2011-10-25T04:18:40 the do_turn? or just the part with for direction in directions etc 2011-10-25T04:18:56 the deafult move if N, E, S or W whichever is passable first 2011-10-25T04:19:22 so in the do_turn func 2011-10-25T04:19:27 *** levis501 has quit IRC (Read error: Connection reset by peer) 2011-10-25T04:19:35 i do orders = {} and then the food stuff? 2011-10-25T04:19:37 does anyone know how to tell currently visible tiles from tiles hidden by the fog of war using the java starter package? 2011-10-25T04:19:37 pvinis: yes 2011-10-25T04:19:47 aha.. 2011-10-25T04:19:51 thanks 2011-10-25T04:20:43 hacklash: that's an excersive for you 2011-10-25T04:20:54 *excercise 2011-10-25T04:21:11 it seems like the starter package should do that but oh well 2011-10-25T04:21:39 hacklash: read the starter package's code, everything will seem simpler 2011-10-25T04:21:57 the starter pack for ruby is just empty 2011-10-25T04:22:10 it only has the n,e,s,w move 2011-10-25T04:22:25 Look at the python starter package pvinis 2011-10-25T04:22:46 *** bearoff has left #aichallenge 2011-10-25T04:22:50 python or python 3? 2011-10-25T04:22:56 Python 2011-10-25T04:23:10 hacklash: I've checked the repo and there is already a isVisible() method in Ants.java 2011-10-25T04:23:48 smoke_carrot: i see what's happening, i assumed the map[][] elements were being left null if not visible, but they are assigned land in the ants constructor and then later reassigned if visible 2011-10-25T04:24:36 sir_macelon: mine doesn't, time to pull a new version 2011-10-25T04:27:27 *** Palmik has quit IRC (Ping timeout: 248 seconds) 2011-10-25T04:27:29 *** thestinger has quit IRC (Quit: WeeChat 0.3.6) 2011-10-25T04:28:17 *** thestinger has joined #aichallenge 2011-10-25T04:31:24 *** sigh has joined #aichallenge 2011-10-25T04:32:48 *** kaemo has quit IRC (Read error: Operation timed out) 2011-10-25T04:38:58 *** porco has joined #aichallenge 2011-10-25T04:40:13 *** the-mgt_ has joined #aichallenge 2011-10-25T04:40:36 *** `mark`` <`mark``!c3711946@gateway/web/freenode/ip.195.113.25.70> has joined #aichallenge 2011-10-25T04:42:04 *** the-mgt has quit IRC (Ping timeout: 258 seconds) 2011-10-25T04:42:04 *** the-mgt_ is now known as the-mgt 2011-10-25T04:48:13 *** randll has quit IRC (Quit: Page closed) 2011-10-25T05:05:00 Rank stabilized seems to happens to early, when a team still has a chance to tie the game. 2011-10-25T05:05:16 *** ashoka has joined #aichallenge 2011-10-25T05:07:50 *** jzknight has quit IRC (Ping timeout: 276 seconds) 2011-10-25T05:08:51 *** jzknight has joined #aichallenge 2011-10-25T05:11:40 *** amstan has quit IRC (Ping timeout: 260 seconds) 2011-10-25T05:12:15 *** Fandekasp has quit IRC (Ping timeout: 260 seconds) 2011-10-25T05:13:10 *** Fandekasp has joined #aichallenge 2011-10-25T05:13:40 *** vasya has quit IRC (Quit: Page closed) 2011-10-25T05:21:02 *** the-mgt has quit IRC (Quit: the-mgt) 2011-10-25T05:22:40 *** xdccFriend_[5260 has joined #aichallenge 2011-10-25T05:22:40 « Un saluto a tutto il canale! » 2011-10-25T05:22:47 *** xdccFriend_[5260 has left #aichallenge 2011-10-25T05:23:57 *** `mark`` <`mark``!c3711946@gateway/web/freenode/ip.195.113.25.70> has quit IRC (Quit: Page closed) 2011-10-25T05:28:00 *** ashoka has quit IRC (Quit: Page closed) 2011-10-25T05:28:51 *** boegel has joined #aichallenge 2011-10-25T05:31:46 *** Draivin has joined #aichallenge 2011-10-25T05:37:55 *** Fandekasp has quit IRC (Ping timeout: 252 seconds) 2011-10-25T05:38:47 oh yeh ml-class homework isn't due till next week. ignore for now XD 2011-10-25T05:38:54 *** pvarga_ has joined #aichallenge 2011-10-25T05:39:36 *** pvarga has quit IRC (Ping timeout: 240 seconds) 2011-10-25T05:39:36 *** pvarga_ is now known as pvarga 2011-10-25T05:40:34 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-10-25T05:41:32 *** master_ninja_ has joined #aichallenge 2011-10-25T05:42:21 *** replore_ has quit IRC (Remote host closed the connection) 2011-10-25T05:45:46 *** asankarc has joined #aichallenge 2011-10-25T05:46:00 *** master_ninja_ has quit IRC (Ping timeout: 265 seconds) 2011-10-25T05:59:03 *** asankarc has quit IRC (Quit: Page closed) 2011-10-25T06:01:09 *** Draivin has quit IRC (Quit: Page closed) 2011-10-25T06:02:42 *** racko has joined #aichallenge 2011-10-25T06:03:22 *** doskir has joined #aichallenge 2011-10-25T06:03:44 What are you guys using to get faster feedback on your bots? I know there are some tcpservers available, and I know some people are duplicating the contest site. I'd just like to know what people actually use :) 2011-10-25T06:04:43 Well, for people like sir_macelon, they simply ~know~ what needs to be improved with their bot. They are simply superhumans. 2011-10-25T06:04:53 *** deltree_ has quit IRC (Ping timeout: 252 seconds) 2011-10-25T06:05:07 So, my advice, beg to the Gods for demigod powers and hope they have mercy. 2011-10-25T06:05:08 How do tcpservers give you faster feedback? 2011-10-25T06:07:09 SharkMonkey: more control over when you will play, no need to upload to the official site, more games 2011-10-25T06:08:40 How is that better than running locally though? 2011-10-25T06:08:53 with playgame.py 2011-10-25T06:08:55 SharkMonkey: you get to play against other people's bots 2011-10-25T06:09:17 Oh. Without revealing your source to them? 2011-10-25T06:09:31 Where are these tcp servers then? 2011-10-25T06:10:36 SharkMonkey: yeah, without revealing your source. someone else will need to chip in on the actual tcp servers, i haven't started participating yet :) 2011-10-25T06:11:54 Thanks for clearing the idea up for me. I'm dense. 2011-10-25T06:14:32 SharkMonkey: you don't reveal your code, since it runs on your machine. 2011-10-25T06:15:25 *** JamesMG_ has joined #aichallenge 2011-10-25T06:17:57 it simply sends the inputs to your bot 2011-10-25T06:18:07 and sends your outputs to other bots 2011-10-25T06:18:33 *** JamesMG has quit IRC (Ping timeout: 256 seconds) 2011-10-25T06:20:23 *** g0llum has joined #aichallenge 2011-10-25T06:20:27 *** JamesMG has joined #aichallenge 2011-10-25T06:20:50 *** Fandekasp has joined #aichallenge 2011-10-25T06:23:12 *** JamesMG_ has quit IRC (Ping timeout: 259 seconds) 2011-10-25T06:25:22 *** thestinger has quit IRC (Quit: WeeChat 0.3.6) 2011-10-25T06:32:08 *** randll has joined #aichallenge 2011-10-25T06:34:09 hi guys! where i can find info about how is skill computed? 2011-10-25T06:35:15 Can we contribute to the FAQ? 2011-10-25T06:35:21 *** alc has quit IRC (Ping timeout: 258 seconds) 2011-10-25T06:35:41 randll: wikipedia TrueSkill 2011-10-25T06:36:45 Garf: thanks a lot! 2011-10-25T06:38:47 *** tampe125 has joined #aichallenge 2011-10-25T06:39:08 hi, i have a question 2011-10-25T06:39:31 should i track down dead ants, too? 2011-10-25T06:39:50 depends on how you're keeping track of ants 2011-10-25T06:40:08 if you're deciding the action of each available ant each turn , then you probably don't need to worry about ants that are dead 2011-10-25T06:40:24 ok thanks 2011-10-25T06:40:28 but i think most people are developing ways to keep track of ants across turns, so that they can assign tasks to individual ants 2011-10-25T06:40:31 can I use BSD3-licensed code in my bot? 2011-10-25T06:40:43 in that case, you'll want to know when a ant is dead, so you can reassign task 2011-10-25T06:41:45 e.g. can I use any third-party code at all? 2011-10-25T06:41:52 ok, but the dead array contains the last known position? or where they died? 2011-10-25T06:42:04 should be where they died 2011-10-25T06:42:24 thnkas 2011-10-25T06:42:49 akosch: not sure on the rules regarding that. shouldn't be a problem I think. 2011-10-25T06:43:05 there's only so many ways to do an A* search lol 2011-10-25T06:43:27 bobbydroptable: yeah, the compiler is BSD3 too... so I guess it's ok 2011-10-25T06:44:01 *** olexs has joined #aichallenge 2011-10-25T06:49:23 *** olexs has left #aichallenge 2011-10-25T06:49:41 *** olexs has joined #aichallenge 2011-10-25T06:50:53 *** xathis has joined #aichallenge 2011-10-25T06:56:39 *** cyphase has joined #aichallenge 2011-10-25T06:58:11 *** tampe125 has quit IRC (Quit: Page closed) 2011-10-25T07:04:42 *** tiger_girl has joined #aichallenge 2011-10-25T07:14:21 *** jhawthorn has joined #aichallenge 2011-10-25T07:17:47 *** sur has joined #aichallenge 2011-10-25T07:18:41 when i run test_bot.cmd .. I get output as "Incorrect number of bots for map" 2011-10-25T07:18:58 any help?? 2011-10-25T07:19:16 sur, add --fill to your cmd 2011-10-25T07:20:34 *** bearoff_w has joined #aichallenge 2011-10-25T07:20:46 thanks...Puj .. and 1 more help.. i submitted my editted code but in my profile my skill is showing -0.00... 2011-10-25T07:21:02 *** bearoff_w has left #aichallenge 2011-10-25T07:21:07 Dunno 2011-10-25T07:22:11 did it play any games? 2011-10-25T07:22:33 bam, exam finished :D 2011-10-25T07:22:39 with over 12 hours to spare 2011-10-25T07:23:52 its showing : "Next game should be within 187.6 minutes." .. So can we submit our code within this time span 2011-10-25T07:24:03 sur: if you submit your rank will be reset 2011-10-25T07:24:08 so if you're going ot do it, just do it 2011-10-25T07:24:21 and it will rerank your new bot 2011-10-25T07:24:24 isn't there a tcp server? 2011-10-25T07:24:28 the main server sucks for testing 2011-10-25T07:24:39 well my bot is completely wrecked 2011-10-25T07:27:33 *** randll has quit IRC (Quit: Page closed) 2011-10-25T07:28:43 antimatroid: yes, there's one, google for it. I don't know how well it works. 2011-10-25T07:28:55 smoke_carrot: i didn't mean for me :P 2011-10-25T07:29:02 i have too much work to write a bot atm 2011-10-25T07:30:41 *** `mark <`mark!c371166b@gateway/web/freenode/ip.195.113.22.107> has joined #aichallenge 2011-10-25T07:32:24 *** Akranis has quit IRC (Quit: Lämnar) 2011-10-25T07:32:37 @later tell amstan sorry about lambdabot. i was demonstrating a couple things and intended to kick her out when we were done, but forgot 2011-10-25T07:32:37 jmcarthur: Ready to serve. 2011-10-25T07:33:12 *** xathis has quit IRC (Quit: Page closed) 2011-10-25T07:33:31 antimatroid: writing a simple bot doesn't take that much time ;-) 2011-10-25T07:34:12 i don't write simple bots :P 2011-10-25T07:34:13 *** Draivin has joined #aichallenge 2011-10-25T07:34:30 smoke_carrot: i wrote the c++ starter bot, that was my simple bot :P 2011-10-25T07:36:16 *** keita has joined #aichallenge 2011-10-25T07:37:32 *** Fandekasp has quit IRC (Ping timeout: 260 seconds) 2011-10-25T07:38:46 *** seering has joined #aichallenge 2011-10-25T07:39:30 If a certian hill has been razed, will the engine still give me it's coordinates? 2011-10-25T07:39:36 no 2011-10-25T07:41:38 *** pvinis has quit IRC (Quit: Leaving) 2011-10-25T07:42:05 @rankings 2011-10-25T07:42:07 sir_macelon: Top 10 players: sir_macelon(80.6), GarySWest(79.0), GreenTea(78.0), MomoBot(77.8), Ender(77.7), ThisIsNotABug(77.3), paratrechina(76.5), ChrisH(75.0), xRaider(74.2), ZumZoom(73.9) 2011-10-25T07:42:42 sir_macelon: bragger :P 2011-10-25T07:43:25 antimatroid: just making a snapshot of rankings for historical reasons :) 2011-10-25T07:43:46 antimatroid: and also inviting you to compete :) 2011-10-25T07:43:50 nice one 2011-10-25T07:44:01 "it's for history!" 2011-10-25T07:44:02 :P 2011-10-25T07:44:18 sir_macelon: once i've handed my thesis in, finished exams and done a 50 minute talk i'll start my bot :P 2011-10-25T07:45:17 antimatroid: how many you have left still? 2011-10-25T07:45:47 one 48 hour take home, one 24 hour take home, and one 1 hour test 2011-10-25T07:45:51 the last two are for the same unit 2011-10-25T07:46:31 and how is it time-wise exactly? 2011-10-25T07:46:45 or you start it whenever you are ready? 2011-10-25T07:46:46 i should be all finished by the 10th or so 2011-10-25T07:46:55 i don't know when the talk is yet 2011-10-25T07:46:57 *** UncleVasya has joined #aichallenge 2011-10-25T07:47:22 antimatroid: well, best of luck then 2011-10-25T07:47:32 cheers 2011-10-25T07:47:50 *** UncleVasya has quit IRC (Read error: Connection reset by peer) 2011-10-25T07:47:59 *** bearoff_w has joined #aichallenge 2011-10-25T07:48:42 is resetting skill after reupload a feature? it's quite annoying, at current game speed it takes more than a day to get anywhere near top 2011-10-25T07:49:49 is there a way to play on a random map using playgame.py? 2011-10-25T07:50:58 i agree with olexs, rank should start where you left off. if the bot is worse the matching algorithm will find out 2011-10-25T07:51:18 but presumably it is the same or better 2011-10-25T07:52:58 olexs: use tcp for testing for bot :P 2011-10-25T07:53:15 antimatroid: I do. still, annoying :) 2011-10-25T07:53:26 resetting requires you to see how your bot does against lower ranked bots and makes it harder for people to keep resubmitting every time they lose a game 2011-10-25T07:53:47 to game the rankings before the final 2011-10-25T07:54:11 hm, true, didn't think about possible problems vs low-ranked bots 2011-10-25T07:55:17 *** danielharan has joined #aichallenge 2011-10-25T07:55:44 i think a few people were surprised there when the planet wars final ran 2011-10-25T07:56:04 planet wars was last time's challenge? 2011-10-25T07:56:06 tcp server had mostly good bots so they never played the lower ranked bots in the final months 2011-10-25T07:56:10 mayanks43: yeah 2011-10-25T07:56:19 k 2011-10-25T07:59:36 *** Claquette has joined #aichallenge 2011-10-25T08:00:12 *** Claquette has quit IRC (Client Quit) 2011-10-25T08:00:29 *** mviel has joined #aichallenge 2011-10-25T08:05:06 *** calsmurf2904 has joined #aichallenge 2011-10-25T08:06:48 Hello, I'm just starting out (with AI programming in general) and I'm currently implementing Pathfinding so I can explore maps etc. without getting stuck....the pathfinding works great but with 20+ ants I get a timeout because it takes way too long for it to calculate paths for all those ants 2011-10-25T08:07:25 Could somebody give me some advice to how I can fix this? I'm using A* pathfinding 2011-10-25T08:08:20 I'm currently at version 8 of the AI and all version 8 games (total of 5) have timed out 2011-10-25T08:09:34 *** danielharan has quit IRC (Remote host closed the connection) 2011-10-25T08:10:12 *** danielharan has joined #aichallenge 2011-10-25T08:10:47 calsmurf2904: you could try and cache stuff, tto avoid computing the same path twice 2011-10-25T08:11:28 also: use an effiicient implementation of A* 2011-10-25T08:11:34 Yes, I was thinking about that...I'm currently also trying to force it to queue the ants who need a new path and just calculate 7 paths per turn 2011-10-25T08:11:49 I'm using my own implementation (which ofcourse isn't really efficient) 2011-10-25T08:13:51 calsmurf2904: see http://www.youtube.com/watch?v=3muiVUU0sys&feature=player_profilepage for efficiency tips 2011-10-25T08:14:08 Thanks :) 2011-10-25T08:15:13 *** Savaron has joined #aichallenge 2011-10-25T08:19:13 *** dgroft has quit IRC (Ping timeout: 265 seconds) 2011-10-25T08:19:17 Wait, so basically, the engine will return only the tiles visible, right? So if my ants go out of range of an enemy hill the engine won't return it's coordinates right? 2011-10-25T08:20:19 seering: yes, just as with food 2011-10-25T08:20:20 only if you don't gather any food 2011-10-25T08:20:47 *** liberforce has joined #aichallenge 2011-10-25T08:22:31 *** ikaros has joined #aichallenge 2011-10-25T08:24:47 *** Paul____ has joined #aichallenge 2011-10-25T08:24:54 hi 2011-10-25T08:24:55 hi guys 2011-10-25T08:25:11 i'm trying t get a C bot working 2011-10-25T08:25:29 is there a way to use external libraries like GLib ? 2011-10-25T08:26:05 people, i extract python_starter_package.zip in tools, and it's doesn't work 2011-10-25T08:26:10 I don't really want to write basic data management structures again 2011-10-25T08:26:40 *** xathis has joined #aichallenge 2011-10-25T08:26:57 *** keita has quit IRC (Ping timeout: 265 seconds) 2011-10-25T08:26:59 ants.py in zip is diffrent of ants.py in tools examples 2011-10-25T08:27:54 how can i test my ants local? 2011-10-25T08:31:15 Paul____: did you follow the instructions here: http://aichallenge.org/ants_tutorial.php ? 2011-10-25T08:31:19 the ants.py in tools is for the game engine 2011-10-25T08:31:22 not for your bot 2011-10-25T08:37:20 *** delt0r_ has quit IRC (Ping timeout: 244 seconds) 2011-10-25T08:37:56 *** ajhager has quit IRC (Ping timeout: 260 seconds) 2011-10-25T08:38:09 thanks, i'll try to read tutorial :) 2011-10-25T08:45:42 oh, tutorial is what i must read first! 2011-10-25T08:46:07 i can't figure out why my bot is timing out 2011-10-25T08:47:06 i'm printing out the time left at the very innermost loop and the logging ends with like 760 left but without exiting that loop 2011-10-25T08:49:46 *** delt0r_ has joined #aichallenge 2011-10-25T08:52:22 *** teuneboon has joined #aichallenge 2011-10-25T08:52:29 hi 2011-10-25T08:53:18 anyone knows how to run playgame.py without immidiatly opening the replay html, but still having it create it? 2011-10-25T08:53:42 I use my server to test my bot, that doesn't have a browser obviously, but I still want to see the replay html file 2011-10-25T08:54:20 I thought --nolaunch was the solution, but that prevents the replay html from being created 2011-10-25T08:56:32 and just found the solution myself by looking in the playgame code(use --html in combination with --nolaunch) 2011-10-25T08:56:44 *** danielharan has quit IRC (Remote host closed the connection) 2011-10-25T08:56:58 I've just found there's something wrong with the tutorial code, it doesn't work properly 2011-10-25T08:57:54 the Java or the Python code? 2011-10-25T08:58:22 *** Garf has quit IRC (Quit: Make a new plan, Stan!) 2011-10-25T08:58:39 teuneboon: both 2011-10-25T08:58:45 in the step 2 at turn 5 ant (29,18) is closer (distance=37) to food (35,19) than ant (28,19) (distance=49), yet the latter one is picked 2011-10-25T08:58:53 *** Garf has joined #aichallenge 2011-10-25T08:59:23 I don't know why, I assume the game is played with python code 2011-10-25T08:59:30 hmm, indeed I see 2011-10-25T08:59:36 maybe sorting in python doesn't work correctly 2011-10-25T08:59:52 but I don't know python 2011-10-25T09:00:06 not very likely considering how many use it 2011-10-25T09:00:24 I know a bit of python, but the sorting there is still somewhat of a mistery to me :P 2011-10-25T09:01:04 *** deltree_ has joined #aichallenge 2011-10-25T09:01:07 it should sort on distance if I look at the code 2011-10-25T09:01:10 strange 2011-10-25T09:01:58 I am reviewing McLeopolds java part and it gives me different outcome 2011-10-25T09:02:19 so for Java it works correctly, so I assume python is broken 2011-10-25T09:02:47 or it is not the actual code in the reply... 2011-10-25T09:03:06 maybe run the python code locally? 2011-10-25T09:03:40 what snippet are we talking about? 2011-10-25T09:04:19 step 2 tutorial 2011-10-25T09:04:21 python 2011-10-25T09:04:50 "find close food" ? 2011-10-25T09:05:29 yeah 2011-10-25T09:06:04 I am working on java tutorial and trying to fix this 2011-10-25T09:06:22 teuneboon: I wouldn't go into python now, maybe you could try 2011-10-25T09:06:39 is the problem the tuple unpacking? 2011-10-25T09:06:59 would have to switch computers to do the python thing XD 2011-10-25T09:07:30 or install python here 2011-10-25T09:07:31 :P 2011-10-25T09:08:33 teuneboon: well, I don't care about python anyway :P 2011-10-25T09:08:46 haha, kind of the same here XD 2011-10-25T09:08:56 started making my bot in Java, now converting to javascript 2011-10-25T09:09:05 since I kind of started that starter bot I think I should use it XD 2011-10-25T09:09:31 teuneboon: oh, I thought you were on python, nvm then 2011-10-25T09:10:24 *** jbh has joined #aichallenge 2011-10-25T09:10:27 I'm talking blind here, but could the issue/difference be in the way the sorts break ties? 2011-10-25T09:10:37 wait, that can'e be it 2011-10-25T09:10:43 locations are unique 2011-10-25T09:10:47 * Garf should just stfu 2011-10-25T09:11:28 *** olexs1 has joined #aichallenge 2011-10-25T09:12:29 only thing I see that's possible is that the actual distance code in the python package uses manhattan distance 2011-10-25T09:12:33 but that would be weird 2011-10-25T09:12:41 i think the python code just finds the nearest ant for every food. but if an ant has 2 possibilities it just sends the ant to the first food processed in the loop 2011-10-25T09:12:49 *** vladimirfol has joined #aichallenge 2011-10-25T09:13:10 ikaros, the list is sorted 2011-10-25T09:13:18 so shouldn't it take the closest first? 2011-10-25T09:13:29 *** olexs has quit IRC (Ping timeout: 260 seconds) 2011-10-25T09:13:38 hi 2011-10-25T09:13:46 hello 2011-10-25T09:13:52 that's a good point actually 2011-10-25T09:13:53 can i write my Bot on python & C using ctypes 2011-10-25T09:13:55 well the list for each food is sorted 2011-10-25T09:13:59 you want to use manhatten for path finding, not euclidean 2011-10-25T09:14:01 > 2011-10-25T09:14:20 and you need to make sure you wrap your distance functions 2011-10-25T09:14:36 I know antimatroid 2011-10-25T09:14:43 i will compile C library source to shared objects library (*.so in linux, *.dll in Windows) on my own computer, send the lib to the server and will call functions from the library in python Bot (using module 'ctypes'). 2011-10-25T09:14:43 maybe i get it wrong im no python guy either 2011-10-25T09:14:43 but the starter packages should use euclidean 2011-10-25T09:14:51 well, java tutorial is even more broken, TreeMap was not a good choice to sort distances as it eliminates all other paths with the same distance... 2011-10-25T09:14:53 no they shouldn't 2011-10-25T09:15:06 oh ok XD 2011-10-25T09:15:16 they do however don't they? 2011-10-25T09:15:20 i say that even though i only gave hte euclidean metric in the c++ one 2011-10-25T09:15:28 but i also didn't do anything related to path finding for people there 2011-10-25T09:15:46 *** hacklash has quit IRC (Ping timeout: 265 seconds) 2011-10-25T09:15:57 *** iVo1337^ has joined #aichallenge 2011-10-25T09:16:09 yeah, I converted the distance in the Java starter package to manhattan myself, but I thought that was something people should find out theirselves 2011-10-25T09:16:22 *by themselves 2011-10-25T09:17:02 *** hacklash has joined #aichallenge 2011-10-25T09:17:14 *** bobbydroptable has quit IRC (Ping timeout: 265 seconds) 2011-10-25T09:17:59 *** ajhager has joined #aichallenge 2011-10-25T09:18:19 so, do allbody (who use script lang) not want to rewrite performance critical funcs on C ? 2011-10-25T09:18:41 i don't know about that, i don't mind not giving something to people but giving them something that's wrong is just setting people up for disaster 2011-10-25T09:19:07 ah wait, I know why they use the "wrong one" 2011-10-25T09:19:12 viewdistances 2011-10-25T09:19:15 are also calculated thay way 2011-10-25T09:19:21 aren't they? 2011-10-25T09:19:23 and attack 2011-10-25T09:19:25 and it's easy to see that the two metrics are not equivalence with 2011-10-25T09:19:25 a.* 2011-10-25T09:19:25 .b. 2011-10-25T09:19:25 a and b have the same manhatten distance but not the same euclidean 2011-10-25T09:19:46 to * without wrapping that is 2011-10-25T09:20:21 *** sur has quit IRC (Quit: Page closed) 2011-10-25T09:20:30 *** seering has quit IRC (Quit: Lost terminal) 2011-10-25T09:20:50 I think you need both distance calculations 2011-10-25T09:20:54 cause to see if an enemy is inside your attack range 2011-10-25T09:21:04 you need the eucledian 2011-10-25T09:21:19 *euclidean 2011-10-25T09:23:12 teuneboon: that's a good point, thanks 2011-10-25T09:23:39 *** monkeys has joined #aichallenge 2011-10-25T09:23:56 *** hacklash has quit IRC (Quit: Page closed) 2011-10-25T09:24:19 *** UncleVasya has joined #aichallenge 2011-10-25T09:24:40 *** clckwrk has joined #aichallenge 2011-10-25T09:25:00 *** hacklash has joined #aichallenge 2011-10-25T09:25:46 *** pedrosorio has joined #aichallenge 2011-10-25T09:26:36 *** jbh has quit IRC (Quit: Page closed) 2011-10-25T09:27:08 Can someone help me get a grasp on radius_sq of view and attack, is it fair to think of this as a square with length of sides {attack,view}radius_sq with ant at the center? 2011-10-25T09:27:08 *** blackwol` is now known as blackwolf 2011-10-25T09:28:40 *** aaa has joined #aichallenge 2011-10-25T09:28:42 clckwrk: http://aichallenge.org/specification.php#Bot-Input scroll down a bit to "distance" 2011-10-25T09:29:11 so it's some kind of circle around the ant 2011-10-25T09:29:17 but not a perfect circle ofc 2011-10-25T09:29:24 since it are squares :) 2011-10-25T09:31:27 right, but I'm trying to say scan the area given by the various radii so although this link can show me if something is within range, it's not valid for say looking 2 squares to the west and seeing what's there. Is that right? 2011-10-25T09:32:42 0A programmers 2011-10-25T09:32:43 2011-10-25T09:32:43 0A young programmers began to work online, 2011-10-25T09:32:46 One didn't pay for Internet, and then there were 9. 2011-10-25T09:32:46 2011-10-25T09:32:46 9 young programmers used copies that they made, 2011-10-25T09:32:46 But one was caught by FBI, and then there were 8. 2011-10-25T09:32:46 2011-10-25T09:32:47 8 young programmers discussed about heaven, 2011-10-25T09:32:47 One said «It's Windows 95!», and then there were 7. 2011-10-25T09:32:48 2011-10-25T09:32:48 7 young programmers found bugs they want to fix, 2011-10-25T09:32:49 But one was fixed by the bug, and then there were 6. 2011-10-25T09:32:49 2011-10-25T09:32:50 6 young programmers were testing the hard drive, 2011-10-25T09:32:50 One got the string «Format complete», and then there were 5. 2011-10-25T09:32:51 2011-10-25T09:32:53 I could take sqrt[radius] and know that I'm able to "see" the floor of that result in all directions with ant at the center 2011-10-25T09:34:02 *** hacklash has left #aichallenge 2011-10-25T09:34:31 *** Paul____ has quit IRC (Quit: Page closed) 2011-10-25T09:35:00 *** sunshowers has joined #aichallenge 2011-10-25T09:36:12 *** clckwrk has quit IRC (Quit: Page closed) 2011-10-25T09:36:25 *** hacklash has joined #aichallenge 2011-10-25T09:36:41 *** monkeys has quit IRC (Quit: Leaving.) 2011-10-25T09:37:13 *** tvorryn has joined #aichallenge 2011-10-25T09:37:30 *** hacklash has left #aichallenge 2011-10-25T09:38:20 *** hacklash has joined #aichallenge 2011-10-25T09:38:36 *** teuneboon has quit IRC (Quit: Page closed) 2011-10-25T09:38:55 *** tvorryn has quit IRC (Client Quit) 2011-10-25T09:39:35 *** clckwrk has joined #aichallenge 2011-10-25T09:39:48 *** TTE has joined #aichallenge 2011-10-25T09:41:26 *** pedrosorio has quit IRC (Quit: Page closed) 2011-10-25T09:42:21 *** sunshowers has quit IRC (Quit: Page closed) 2011-10-25T09:50:33 *** Xeor has joined #aichallenge 2011-10-25T09:50:51 *** Xeor is now known as Xeorr 2011-10-25T09:51:03 *** calsmurf2904 has quit IRC (Ping timeout: 265 seconds) 2011-10-25T09:52:50 What value for turntime is used in the tournament? 2011-10-25T09:55:46 antonh: Now it's 500. 2011-10-25T09:56:28 *** olexs1 has quit IRC (Quit: Leaving.) 2011-10-25T09:57:15 *** krikav has joined #aichallenge 2011-10-25T09:58:54 *** Harpyon has joined #aichallenge 2011-10-25T10:01:24 looked for games for just submitted bot version, and find this game: http://aichallenge.org/visualizer.php?game=11161&user=180 What happened with all of them? Server fault? 2011-10-25T10:01:34 antonh: http://aichallenge.org/game_settings.php 2011-10-25T10:01:50 avdg: thanks, cool 2011-10-25T10:02:19 you can get actual settings for game by presiing "bot input" button too 2011-10-25T10:02:30 in vizualizer 2011-10-25T10:02:49 though i wish I don't have to type a username ;-) 2011-10-25T10:03:20 *** emilk has joined #aichallenge 2011-10-25T10:03:26 *** twymer has joined #aichallenge 2011-10-25T10:04:39 if you ca,e to visualiser from player's profile, his name will be pre-entered, it seems 2011-10-25T10:04:43 *came 2011-10-25T10:04:53 *** Saulzar has joined #aichallenge 2011-10-25T10:05:46 *** doskir has left #aichallenge 2011-10-25T10:07:21 bearoff_w that game had a bunch of starter bots which all did the same thing and suicided. no crashes just bad programming :) 2011-10-25T10:08:16 lol mass suicide in the same round :) 2011-10-25T10:09:09 *** Bleda has joined #aichallenge 2011-10-25T10:09:24 *** cbad has quit IRC (Read error: Connection reset by peer) 2011-10-25T10:09:42 oh, didn't notice that they do suicide ) 2011-10-25T10:10:24 it was a good proposition to give only 1 game for just uploaded starter bots - no need to play games with them 2011-10-25T10:11:42 well if you upload a new version you are starting from 0 rating 2011-10-25T10:12:01 it wouldn't matter anyways if there was more server capacity :) 2011-10-25T10:12:41 but in current situation it matters ( 2011-10-25T10:13:16 *** foRei has joined #aichallenge 2011-10-25T10:14:41 yea just use the tcp server for testing and only upload new versions if they are really better 2011-10-25T10:14:58 I just do this ) 2011-10-25T10:15:02 :) 2011-10-25T10:15:14 *** cbad has joined #aichallenge 2011-10-25T10:15:14 *** cbad has joined #aichallenge 2011-10-25T10:15:16 are there any plans to upgrade servers further btw? 2011-10-25T10:15:33 at least there are 5 servers now 2011-10-25T10:15:38 yesterday it was 3 2011-10-25T10:15:45 and 2 at beginning 2011-10-25T10:16:16 i see 5 listed but they have names #7 - #11 2011-10-25T10:16:47 there can be a lot of reasons ) 2011-10-25T10:16:52 =) 2011-10-25T10:17:43 ikaros: every started server gets a new ID, so this is not the actual count 2011-10-25T10:17:58 ok 2011-10-25T10:18:32 *** swamp has joined #aichallenge 2011-10-25T10:21:25 *** he_the_great has joined #aichallenge 2011-10-25T10:21:43 *** swamp has left #aichallenge 2011-10-25T10:22:43 *** sigh has quit IRC (Remote host closed the connection) 2011-10-25T10:23:15 *** sigh has joined #aichallenge 2011-10-25T10:27:49 *** sigh has quit IRC (Ping timeout: 252 seconds) 2011-10-25T10:29:22 Is there a plan to increase the number of servers even more? 2011-10-25T10:31:22 *** vladimirfol has quit IRC (Quit: Page closed) 2011-10-25T10:32:12 hey guys, is it possible to link dinamically to external libs like GLib ? 2011-10-25T10:32:32 does the server have glib installed ? 2011-10-25T10:32:48 *** Savaron has quit IRC (Quit: Page closed) 2011-10-25T10:32:49 or should i link statically for this ? 2011-10-25T10:34:08 i don't think static linkage would work, as the server rebuilds the bots... 2011-10-25T10:34:26 McLeopold: I've finished reviewing and updating the Java tutorial, you can have a look if everything's fine 2011-10-25T10:34:29 but i don't know which OS the server uses 2011-10-25T10:34:33 liberforce: correct it rebuilds 2011-10-25T10:35:41 So C is a bit disadvantaged here, as python and oters have linked lists and other stuff implementations out-of the box 2011-10-25T10:35:51 Oh sweet, I am in the top 100 now! 2011-10-25T10:35:57 nice ;) 2011-10-25T10:35:57 *** Savaron has joined #aichallenge 2011-10-25T10:36:01 and the bot I have up there is an older version 2011-10-25T10:36:17 new version is much better 2011-10-25T10:36:25 after one more update ill upload it again 2011-10-25T10:37:03 althought 3 days for feedback on a bot is pretty silly, the TCP servers are where it is at 2011-10-25T10:38:05 maybe getting amazon as sponsor would have been better than google :P 2011-10-25T10:38:11 (afaik the servers are EC2 instances, right) 2011-10-25T10:38:20 *** twymer has quit IRC (Quit: leaving) 2011-10-25T10:41:16 *** mcstar has joined #aichallenge 2011-10-25T10:41:44 McLeopold: hi 2011-10-25T10:44:41 *** he_the_great has quit IRC (Ping timeout: 252 seconds) 2011-10-25T10:50:47 yay, optimizations. my bot is now able to handle almost 700 ants in less than 400ms per turn 2011-10-25T10:52:00 *** dungeonduke has joined #aichallenge 2011-10-25T10:52:37 *** pvarga has quit IRC (Quit: pvarga) 2011-10-25T10:53:50 *** cyphase has quit IRC (Ping timeout: 260 seconds) 2011-10-25T10:54:09 *** Savaron has quit IRC (Quit: Page closed) 2011-10-25T10:55:29 sir_macelon: Hi. Is berak's server script only for linux? If no, my friend wanna make one more TCP. And is the number of already working TCP-servers enough or more is always better? 2011-10-25T10:56:44 UncleVasya: why do you need more tcp servers? 2011-10-25T10:56:50 tcp servers are for testing purposes 2011-10-25T10:57:02 UncleVasya: I guess the OS doesn't matter as your bot connects to it via TCP. I don't know if it is enough, but I guess every one more is beneficial 2011-10-25T10:57:13 *** cyphase has joined #aichallenge 2011-10-25T10:57:18 I don't need. I wanna to figure out are they needed or not. 2011-10-25T10:57:42 *I don't know if there's enough of those TCP servers 2011-10-25T10:57:52 *** ig has quit IRC (Read error: Connection reset by peer) 2011-10-25T10:58:17 More TCP servers just spreads out the players more, making matchmaking harder, I'd think. 2011-10-25T10:58:26 thank you, I'll try to remember. 2011-10-25T10:58:39 cerr(fmt.Sprintf("%d : %4d (%d)\n", s.turn, i, dt)) 2011-10-25T10:58:40 it's not like adding more tcp servers helps real contest 2011-10-25T10:58:40 Could be a point if the existing servers are all overloaded (I have no idea) 2011-10-25T10:59:05 test on tcp, if it is good, upload to official server 2011-10-25T10:59:07 can two ants swap places in one turn? like ab -> ba 2011-10-25T10:59:12 ikaros: yes 2011-10-25T10:59:16 ok good 2011-10-25T10:59:26 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-10-25T10:59:41 *** ig has joined #aichallenge 2011-10-25T10:59:50 *** ai4 has joined #aichallenge 2011-10-25T11:04:22 *** onensora has quit IRC () 2011-10-25T11:07:05 *** sir_macelon has quit IRC (Quit: Page closed) 2011-10-25T11:07:16 *** aerique has quit IRC (Quit: ...) 2011-10-25T11:10:23 how is everyone profiling python bots? just cProfile.run('main()', 'output') with a greater load/turn time? 2011-10-25T11:13:06 *** racko has quit IRC () 2011-10-25T11:13:39 McLeopold suggested I run the playgame.py with -m cProfile 2011-10-25T11:14:04 that would profile playgame.py 2011-10-25T11:14:37 you could pass python -m cProfile -o output_file YourBot.py as the command to run 2011-10-25T11:14:53 It also records profiling for MyBot I believe. At least some of the methods are included. My bot doesn't not suck enough to confirm the method though. 2011-10-25T11:15:34 i'm not sure why it would. the bots are run as a subprocess, not by importing it as part of playgame.py 2011-10-25T11:16:02 I agree with you. 2011-10-25T11:17:09 Okay so you don't get your method timings, but you do get how much time is spent in your bot.. in total 2011-10-25T11:17:10 1 0.000 0.000 0.000 0.000 MyBot.py:12(MyBot) 1 0.000 0.000 0.000 0.000 MyBot.py:13(__init__) 1 0.058 0.058 7.317 7.317 MyBot.py:2() 2011-10-25T11:17:12 Sorry 2011-10-25T11:17:16 *** g0llum has joined #aichallenge 2011-10-25T11:17:19 1 0.058 0.058 7.317 7.317 MyBot.py:2() 2011-10-25T11:17:23 i have MyBot.sh and ( time python3 -mcProfile -o ./lol.profile ./src/MyBot.py3 ) 2> ./log.txt in it 2011-10-25T11:18:19 *** `mark <`mark!c371166b@gateway/web/freenode/ip.195.113.22.107> has quit IRC (Quit: Page closed) 2011-10-25T11:18:43 And you use MyBot.sh in your play_one_game, for example, Fluxid? 2011-10-25T11:18:49 Or where do you get your input? 2011-10-25T11:18:59 Puj: yes 2011-10-25T11:19:08 Damn, that's a sweet solution. May I? 2011-10-25T11:19:15 sure, go on 2011-10-25T11:19:18 tyty 2011-10-25T11:19:31 this redirects stderr of time and bot to file 2011-10-25T11:21:04 Do you need the parentheses in there because it is being inlined into the other shell script? 2011-10-25T11:21:28 Or just to denote which stderr to steal? 2011-10-25T11:22:10 i need () to redirect time output to file too 2011-10-25T11:22:18 without () only stderr of bot goes to file 2011-10-25T11:22:27 Ah, right o. 2011-10-25T11:23:32 *** boegel has quit IRC (Quit: *poof!*) 2011-10-25T11:25:18 *** tiger_girl has quit IRC (Ping timeout: 265 seconds) 2011-10-25T11:25:19 *** sunshowers has joined #aichallenge 2011-10-25T11:28:41 *** sur has joined #aichallenge 2011-10-25T11:29:01 *** Areks has quit IRC (Ping timeout: 240 seconds) 2011-10-25T11:29:18 how do we specify no. of bots while executing testbot.py 2011-10-25T11:29:24 *** olexs has joined #aichallenge 2011-10-25T11:29:26 *** lavalamp has joined #aichallenge 2011-10-25T11:29:27 *** SharkMonkey has quit IRC (Ping timeout: 260 seconds) 2011-10-25T11:29:53 *** Nescio has joined #aichallenge 2011-10-25T11:30:17 If I register an account and later decide that I want to form a team (the F.A.Q. says teams are ok) can I delete my first account so I don't break the single account rule? 2011-10-25T11:31:20 ?? any help in how to specify no. of bots in terminal 2011-10-25T11:32:18 hi 2011-10-25T11:34:23 Fluxid, can I do this "p_01.map" %* "%~dp0MyBot.sh"" 2011-10-25T11:35:29 *** Racko has joined #aichallenge 2011-10-25T11:35:58 when i run test_map.cmd i get incorrect no. of bots . Need 2 got 1. can u say how to specify no. of bots... 2011-10-25T11:37:03 aichallenge: sir-macelon epsilon * rb89fd1b / ants/dist/starter_bots/java/Tile.java : Make Tile class comparable. - http://git.io/bayp0Q 2011-10-25T11:37:04 aichallenge: Scott Hamilton epsilon * r5474e40 / ants/dist/starter_bots/java/Tile.java : 2011-10-25T11:37:04 aichallenge: Merge pull request #331 from sir-macelon/patch-2 2011-10-25T11:37:04 aichallenge: Make Tile class comparable. - http://git.io/kEUN6Q 2011-10-25T11:37:51 comparing by hashcode? interesting xD 2011-10-25T11:38:37 hm, nevermind, the way hashcode is implemented it even sort of makes sense 2011-10-25T11:39:16 *** cooke has joined #aichallenge 2011-10-25T11:39:30 sur: Add their names like p_01.map... MyBot.exe MyBot2.exe 2011-10-25T11:39:50 sur: use --fill 2011-10-25T11:42:31 @avdg when we use test_bot.cmd --fill 2 will this test my new code Mybot.java and can we see the map in browser 2011-10-25T11:42:32 sur: An error occured while trying to show the previous error. 2011-10-25T11:42:57 *** ig has quit IRC (Read error: Connection reset by peer) 2011-10-25T11:43:12 @avdg when we use test_bot.cmd --fill 2 will this test my new code Mybot.java and can we see the map in browser 2011-10-25T11:43:13 sur: User error, it's not my fault. 2011-10-25T11:43:23 sur: don't use the @ 2011-10-25T11:43:23 --fill doesn't take a parameter 2011-10-25T11:43:26 when we use test_bot.cmd --fill 2 will this test my new code Mybot.java and can we see the map in browser 2011-10-25T11:43:45 *** hacker007 has joined #aichallenge 2011-10-25T11:43:51 Sur as in Swedish for sour or Sur as in Spanish for south? 2011-10-25T11:44:17 Puj nothing like that... Indian 2011-10-25T11:44:23 *** ig has joined #aichallenge 2011-10-25T11:45:59 *** chaudeau has joined #aichallenge 2011-10-25T11:46:52 meh, my bot played 3 times at almost the same time 2011-10-25T11:47:32 *** Racko has quit IRC (Ping timeout: 265 seconds) 2011-10-25T11:47:45 *** cooke has quit IRC (Quit: Page closed) 2011-10-25T11:47:48 avdg: it's weird, it seems when your bot is near top of the queue it can get chosen for a game multiple times, and then goes back to the queue end 2011-10-25T11:48:19 yeah, it was not the first time 2011-10-25T11:48:34 but 3 times? that gets my attention 2011-10-25T11:48:49 olexs: that is probably because we are in a section of bots that are near your rank 2011-10-25T11:49:42 McLeopold: matchmaking always takes the first player in queue, and then finds match partners for him further up the queue, right? 2011-10-25T11:49:55 somewhat, yes 2011-10-25T11:50:39 further up the queue is not the actual logic, we look for the best trueskill match quality within a range of players 2011-10-25T11:50:58 okay 2011-10-25T11:51:03 it is a pareto distribution, so you have an 80% chance of playing the 16 closest players 2011-10-25T11:51:18 16 might be off, I can't remember where we ended 2011-10-25T11:51:31 * avdg wants graphs :p 2011-10-25T11:51:37 what kind? 2011-10-25T11:51:50 ah http://www.wolframalpha.com/input/?i=pareto+distribution 2011-10-25T11:52:25 *** Cyndre_ is now known as Cyndre 2011-10-25T11:53:25 https://github.com/aichallenge/aichallenge/blob/epsilon/sql/2_generate_matchup.sql#L251 there's the sql code to generate the range 2011-10-25T11:53:31 *** skalas has joined #aichallenge 2011-10-25T11:53:38 *** dauryg has joined #aichallenge 2011-10-25T11:53:39 i restarted the tcp server, it kind of hung with one game and 11 connected players... 2011-10-25T11:54:00 one of my favourite wa question: "are you skynet?" 2011-10-25T11:54:18 Yo Fluxid I'm getting SandboxError: Failed to start ['./MyBot.sh'] 2011-10-25T11:54:30 But I'm in the tools directory where the shell script is. 2011-10-25T11:54:34 And it has rights. 2011-10-25T11:55:06 *** the-mgt has joined #aichallenge 2011-10-25T11:56:11 Puj: try absolute path 2011-10-25T12:01:06 *** JamesMG_ has joined #aichallenge 2011-10-25T12:04:22 k thanks 2011-10-25T12:04:38 *** JamesMG has quit IRC (Ping timeout: 258 seconds) 2011-10-25T12:06:21 http://ants.fluxid.pl/replay.5459 haha, i like how my bot behaves from turn 400 on 2011-10-25T12:08:04 Fluxid, hahah :D 2011-10-25T12:10:25 how can we see our code running in maps ie if we run in terminal test_bot.cmd --fill we dont see the maps 2011-10-25T12:10:26 nice 2011-10-25T12:11:31 sur: java -jar visualizer.jar ./game_logs/0.replay 2011-10-25T12:11:49 if you are in ants folder 2011-10-25T12:11:51 *** dungeonduke has quit IRC (Remote host closed the connection) 2011-10-25T12:12:03 and you have compiled visualizer 2011-10-25T12:14:36 *** chaudeau has quit IRC (Ping timeout: 265 seconds) 2011-10-25T12:14:42 *** iris1 has quit IRC (Read error: Connection reset by peer) 2011-10-25T12:16:05 *** onensora has joined #aichallenge 2011-10-25T12:16:17 *** hacker007 has quit IRC (Quit: Page closed) 2011-10-25T12:16:58 *** exobit__ has joined #aichallenge 2011-10-25T12:21:04 Does anybody know how big the fog of war is around each ant? 2011-10-25T12:21:44 exobit__, viewradius2 is the square of the view radius 2011-10-25T12:21:48 *** Areks|2 has joined #aichallenge 2011-10-25T12:23:04 *** sunshowers has quit IRC (Quit: Page closed) 2011-10-25T12:23:17 cyphase: Does it change depending on the map? 2011-10-25T12:23:35 exobit__, i don't think so, but don't quote me on that 2011-10-25T12:24:00 exobit__: no, but may change anyway, so don't rely on constant value 2011-10-25T12:24:02 *** amstan has joined #aichallenge 2011-10-25T12:24:02 *** ChanServ sets mode: +o amstan 2011-10-25T12:24:03 contestbot: seen bearoff 2011-10-25T12:24:04 amstan: bearoff was last seen in #aichallenge 17 hours, 47 minutes, and 30 seconds ago: on main board is old stupid version, if interesting ) 2011-10-25T12:25:07 *** tncardoso has joined #aichallenge 2011-10-25T12:27:54 *** Blkt has quit IRC (Read error: Connection reset by peer) 2011-10-25T12:28:16 amstan, wrt. the ants-tcp thread on the forums: i still think you should set up an official tcp server at some point and direct people to it :P 2011-10-25T12:28:32 Alexer: we never had an official tcp server 2011-10-25T12:28:55 amstan, yes, but it'd be cool if you did 2011-10-25T12:29:14 it's extra stuff we have to do, we'd rather work on making the next contest better, or fixing current issues with the current contest 2011-10-25T12:29:37 every contest there was someone else with the tcp project, they usually handled it well 2011-10-25T12:30:28 amstan, well it's just that there are 3 tcp servers currently, and some people have mentioned wanting to host another one, and that's just dividing the player base 2011-10-25T12:30:40 so what's the problem? 2011-10-25T12:30:41 it'd be nice if there was an official server to direct people to.. 2011-10-25T12:31:00 nobody says there shall only be one server 2011-10-25T12:31:04 amstan: Are there already some thoughts about next contest? Will it be zeta or another one-game? Will you think about it before this one ends or after that? 2011-10-25T12:31:26 UncleVasya: the codebase is going to be called beta, we'll probably get django for the website 2011-10-25T12:31:33 UncleVasya: we don't have a game for it yet though 2011-10-25T12:31:36 zeta* 2011-10-25T12:31:48 what about my awesome nebular wars! 2011-10-25T12:31:50 :D 2011-10-25T12:31:55 Its nebular! 2011-10-25T12:31:58 delt0r: what about it? 2011-10-25T12:32:05 delt0r: i haven't seen it 2011-10-25T12:32:07 *** mviel has quit IRC (Remote host closed the connection) 2011-10-25T12:32:16 well you know... its like would be cool for the next comp 2011-10-25T12:32:32 do you have a description somewhere? 2011-10-25T12:32:42 well no i kinda didn't finish it when it was clear ants was picked 2011-10-25T12:33:02 but yea somewhere on the forums 2011-10-25T12:33:24 Or turn based "squad command" 2011-10-25T12:33:40 something like that i would put more time into 2011-10-25T12:33:42 amstan, but having more than one server does no good, it only makes it harder 2011-10-25T12:33:51 well i would if i had any ;) 2011-10-25T12:33:54 i would like something like ants, but with changable terrain 2011-10-25T12:34:02 *** Knekkebjoern has quit IRC (Quit: Leaving.) 2011-10-25T12:34:18 not a fan 2011-10-25T12:34:38 changeable terrian just complicated the bots without much gain IMHO 2011-10-25T12:34:47 complicates 2011-10-25T12:35:01 currently fluxids server seems to be the de facto server though. ashods server has no players and accouns server has maybe two or three, whereas fluxids server has more 2011-10-25T12:35:47 Alexer: so get fluxid to post a topic on the forums, and i'll replace accoun's post 2011-10-25T12:36:05 ... 2011-10-25T12:36:08 amstan: why replace... sounds like one more flame war 2011-10-25T12:36:17 all coordinates are 0-indexed (row,col), right? 2011-10-25T12:36:29 yep 2011-10-25T12:36:47 delt0r: yeah, maybe 2011-10-25T12:36:58 delt0r, because Accouns post doesn't mention any server besides his own, which doesn't have enough players to play 2011-10-25T12:37:11 amstan: post about what? tcp? and what should i tell? 2011-10-25T12:37:20 Alexer: So? just add a post... 2011-10-25T12:37:26 don't think replacing is a good idea 2011-10-25T12:37:31 Fluxid: introduce your server, give a short tutorial on what to download and how you should start compeeting 2011-10-25T12:37:39 Fluxid: maybe include a link to the server code 2011-10-25T12:37:49 delt0r, i asked Accoun to add the other servers to the first post, but he didn't. the servers are mentioned in the middle of the second page... 2011-10-25T12:37:50 delt0r: posts get lost easily 2011-10-25T12:38:08 delt0r: crucial info is usually on the first post 2011-10-25T12:38:21 Alexer: yea we know... there is history here, and it is likely to cause more issues that it will save 2011-10-25T12:38:22 amstan: i'm hosting this server for testing purposes, to play with other bots before sending my bot to official server. it's not for competition. 2011-10-25T12:38:34 if someone wants to compete on my server.. he got the wrong idea 2011-10-25T12:38:52 also dont think deleting/replacing posts in forums is a good idea generally 2011-10-25T12:38:56 Fluxid, no-ones saying it's for competition 2011-10-25T12:39:04 and code is the same from github as in http://aichallenge.org/forums/viewtopic.php?f=25&t=1505 2011-10-25T12:39:30 Fluxid, i'm just saying yours is currently the only one you can test your bot on, since the others haven't got enough players 2011-10-25T12:39:37 *** skunx has quit IRC (Quit: leaving) 2011-10-25T12:39:55 Alexer: ok 2011-10-25T12:39:57 Fluxid: it's for testing, of course 2011-10-25T12:40:03 Fluxid: you could even say that in your post 2011-10-25T12:40:15 *** Puj has quit IRC (Ping timeout: 260 seconds) 2011-10-25T12:40:24 lol, the example for turn 1 input doesn't include the players own hill: http://aichallenge.org/specification.php#Bot-Input 2011-10-25T12:40:36 cyphase: fix it 2011-10-25T12:40:49 Fluxid, how heavy is the server btw? 2011-10-25T12:41:02 an idea for the next contest: write an AI that succesfully completes a jump map on urban terror 2011-10-25T12:41:14 Alexer: not at all. 2011-10-25T12:41:34 *** FLashM has joined #aichallenge 2011-10-25T12:41:47 Fluxid, it's not the same code either, btw. if you look at accouns server, the tcpclient.py there hasn't been updated, for example. so it won't work as-is for linux users.. 2011-10-25T12:42:09 i've pulled changes before, should i pull again? 2011-10-25T12:42:16 (it works for me) 2011-10-25T12:42:35 Fluxid, no, i meant you have updated, accoun has not :) 2011-10-25T12:42:49 mcstar: games would be too cpu intensive 2011-10-25T12:43:01 *** ig has quit IRC (Read error: Connection reset by peer) 2011-10-25T12:43:36 *** ig has joined #aichallenge 2011-10-25T12:44:29 okay, newest version is not as good but comparable, gonna upload it 2011-10-25T12:45:00 Fluxid, huh? 2011-10-25T12:45:21 of my bot 2011-10-25T12:45:25 sorry, talking to self 2011-10-25T12:45:26 ah 2011-10-25T12:45:36 :) 2011-10-25T12:46:24 amstan, what do you mean? 2011-10-25T12:46:40 cyphase: you found an error, fix it 2011-10-25T12:46:48 amstan, just say if i'm peing too pushy, but i'm going to try to justify it one more time :P 2011-10-25T12:47:08 amstan, how exactly am i supposed to do that? 2011-10-25T12:47:49 cyphase: edit this: https://github.com/aichallenge/aichallenge/wiki/Ants-Bot-Input 2011-10-25T12:47:57 aah, there it is 2011-10-25T12:48:03 i couldnt find it in the repo 2011-10-25T12:49:19 *** shark_teef has joined #aichallenge 2011-10-25T12:49:50 *** freopen1 has joined #aichallenge 2011-10-25T12:49:59 amstan, since it takes a lot more time than what would be nice to test your bot on the normal official server, i think it would be beneficial to direct people to test their bots on a tcp server, since it's much faster 2011-10-25T12:51:02 amstan, currently people often come in here, say something like "next game in 200 minutes, for real??", and we direct them to test on the tcp servers 2011-10-25T12:51:19 *** durarara has joined #aichallenge 2011-10-25T12:51:28 *** durarara is now known as aliceinwire 2011-10-25T12:51:41 hello 2011-10-25T12:51:43 Rank: 2326 2011-10-25T12:51:44 lol 2011-10-25T12:51:52 so it'd be cool if the info was on the site itself. however i think someone from the staff said at one point that you can't really tell people to go play on a server you don't host yourself 2011-10-25T12:52:01 and Fluxid might not like that either >:) 2011-10-25T12:52:02 *** ksteele has joined #aichallenge 2011-10-25T12:52:03 next game in one hour, good 2011-10-25T12:52:04 Alexer: that 2011-10-25T12:52:15 i have finished the tutorial but why my ants after some turn get freeze and they dont move? 2011-10-25T12:52:41 usually after 200 move they freeze 2011-10-25T12:53:37 amstan, so i guess my point is that if you had an official tcp server, you could tell people on the site to test on that if they can, since it's much faster 2011-10-25T12:53:40 *** Macuyiko has joined #aichallenge 2011-10-25T12:53:45 *** freopen1 has quit IRC (Client Quit) 2011-10-25T12:54:16 @later tell Smiley1983 Hey! I have fixed a problem with gettimeofday function by replacing Ants.ml MyBot.ml Unix.cmxa -o MyBot.exe with Unix.cmxa Ants.ml MyBot.ml -o MyBot.exe . Heheh :D 2011-10-25T12:54:16 UncleVasya: Aye, aye, sir 2011-10-25T12:54:35 *** KennyS has joined #aichallenge 2011-10-25T12:54:51 *** bearoff_w has left #aichallenge 2011-10-25T12:55:12 amstan, since waiting for a day to see how your changes affect your play against the other players isn't really nice 2011-10-25T12:55:32 *** KennyS has quit IRC (Client Quit) 2011-10-25T12:56:41 Alexer: they do their best but this all about money. 2011-10-25T12:56:56 *it's all 2011-10-25T12:56:59 I hope 2011-10-25T12:57:04 any help ? 2011-10-25T12:57:08 UncleVasya: money? 2011-10-25T12:57:11 aliceinwire: you mean timeouts? 2011-10-25T12:57:31 UncleVasya, if they tell me it's about money, i'll listen, but no-one has told me so 2011-10-25T12:57:38 amstan: if i reupload the code, i get matches with bots from lowest rank, right? so i need to climb all the way up to place 30 once more? 2011-10-25T12:57:47 Fluxid: i'm not sure 2011-10-25T12:57:51 I hope you don't mind I ask again, but if I register an account and later decide that I want to form a team (the F.A.Q. says teams are ok) can I delete my first account so I don't break the single account rule? 2011-10-25T12:57:55 Fluxid, yes 2011-10-25T12:58:07 Fluxid, more or less 2011-10-25T12:58:10 jix_: sure, can you? 2011-10-25T12:58:38 amstan: When I asked you about will you try to keep some rate you said it depends on money. Do I misunderstand you? 2011-10-25T12:58:54 amstan: That's what I'm asking (I havent registered yet) 2011-10-25T12:58:58 *** Knekkebjoern has joined #aichallenge 2011-10-25T12:59:02 *** cybsy has joined #aichallenge 2011-10-25T12:59:02 jix_: yep 2011-10-25T12:59:13 ok, thanks 2011-10-25T12:59:22 It is quite painful to wait for an effective result from reuploaded bot, since it takes up to 12 hours to make it into the middle of the ranking table 2011-10-25T12:59:33 UncleVasya: yes, game rates are dependent on money, but that doesn't exclude a tcp server 2011-10-25T12:59:39 amstan: yes timeout 2011-10-25T12:59:52 aliceinwire: that means your bot took too long to do its thing 2011-10-25T12:59:57 And I for one can't beleive that google couldn't add several servers to speed up the process 2011-10-25T13:00:00 I told nothing about tcp. 2011-10-25T13:00:22 *** deltree__ has joined #aichallenge 2011-10-25T13:00:44 Fluxid, at least it resets your skill to 0, and it gradually starts to increase, which takes a lot of time in the start when you have low skill, later it seems to go much faster 2011-10-25T13:00:46 I just say that low game rate is not a thing organisers must be blamed for. 2011-10-25T13:02:00 *** deltree_ has quit IRC (Ping timeout: 256 seconds) 2011-10-25T13:02:06 UncleVasya, i'm not blaming anyone :) i'm just saying there is a way to provide a *lot* higher game rate easily, for those that want to use it for testing 2011-10-25T13:02:28 sure, you can use it already, but it's harder to endorse it when it's not official 2011-10-25T13:02:39 Is there any kind of howto with public tcp server adresses and instructions? 2011-10-25T13:02:42 amstan: so make the code in c can be more faster ? 2011-10-25T13:02:54 on doing things? 2011-10-25T13:03:01 aliceinwire: it's not necessarly on language speed 2011-10-25T13:03:10 aliceinwire: you might have a similar problem in C 2011-10-25T13:03:20 FLashM, http://ants.fluxid.pl/howto 2011-10-25T13:03:33 aliceinwire: you need to figure out how to do your algorithms to be faster 2011-10-25T13:03:33 Thanks a lot 2011-10-25T13:03:38 Alexer: ah, I haven't saw your very first message. I apologize. 2011-10-25T13:03:45 aliceinwire: or if you're doing searching, don't do it for all ants for long paths 2011-10-25T13:03:52 UncleVasya, no problem :) 2011-10-25T13:04:23 *** dvladim has joined #aichallenge 2011-10-25T13:05:44 amstan: thanks you :) 2011-10-25T13:06:10 Fluxid, how much time it took you to set up the tcp server? 2011-10-25T13:06:19 Alexer: five minutes 2011-10-25T13:06:24 if not less 2011-10-25T13:06:56 nginx vhost with proxy, start one process, edit one file, start second process, done 2011-10-25T13:07:26 amstan, so sure, it'd be extra stuff, but apparently it wouldn't take long, at least :P 2011-10-25T13:08:38 *** danielharan has joined #aichallenge 2011-10-25T13:09:12 *** skalas has quit IRC (Quit: Leaving) 2011-10-25T13:10:17 for should have a 4th part: for(int i = 0; i < 10 ? i != 0 ? cout << "," : true : false ; i++) 2011-10-25T13:10:53 "execute this, except for the last iteration" 2011-10-25T13:11:00 *** shifu has joined #aichallenge 2011-10-25T13:15:15 my bot executes in less than .5 seconds locally but times out on the server. Is there any other failure condition that could result in a timeout message? http://aichallenge.org/profile.php?user=2238 2011-10-25T13:15:46 lavalamp, maybe you're giving it more processing power than the server 2011-10-25T13:16:23 It's running in a VM with limited RAM on my home computer, so it's possible but it seems unlikely. 2011-10-25T13:16:28 *** exobit__ has quit IRC (Ping timeout: 265 seconds) 2011-10-25T13:17:02 do you have code to check the time remaining in the turn? 2011-10-25T13:17:47 lavalamp: maybe you test only on several maps and the bug in your code doesn't do it's job on them. Try to test bot on many maps and launch it against himself on all of them. 2011-10-25T13:18:24 *** amstan has quit IRC (Ping timeout: 240 seconds) 2011-10-25T13:18:24 *** ksteele has quit IRC (Ping timeout: 265 seconds) 2011-10-25T13:18:53 cyphase: no, because it's barely doing anything. 2011-10-25T13:19:12 UncleVasya: Yeah, I haven't done extensive testing. 2011-10-25T13:19:35 btw, for the next contest I think it's time to step away from the battles, conquests, deaths and blood. And do something in sport/racing/card settings. 2011-10-25T13:19:51 lavalamp: Are you under Windows? 2011-10-25T13:20:08 *setting 2011-10-25T13:21:48 *** fap-machine has joined #aichallenge 2011-10-25T13:22:14 hi 2011-10-25T13:23:28 *** lvqinghouer has joined #aichallenge 2011-10-25T13:23:40 hi 2011-10-25T13:26:11 hm, UncleVasya welcomes fap-machines openly 2011-10-25T13:27:00 *** hacklash has quit IRC (Quit: hacklash) 2011-10-25T13:27:09 *** Racko has joined #aichallenge 2011-10-25T13:29:52 *** chaudeau has joined #aichallenge 2011-10-25T13:32:17 *** Redgis has joined #aichallenge 2011-10-25T13:34:23 *** foRei has quit IRC (Read error: Connection reset by peer) 2011-10-25T13:35:00 *** liberforce has left #aichallenge 2011-10-25T13:36:07 *** ai4 has quit IRC (Quit: Page closed) 2011-10-25T13:36:36 UncleVasya: My VM is ubuntu, host is windows. 2011-10-25T13:37:15 *** FLashM has quit IRC (Ping timeout: 265 seconds) 2011-10-25T13:38:01 mcstar: Can't understand this your joke. Can you explaine? 2011-10-25T13:38:28 lavalamp: never mind 2011-10-25T13:39:45 UncleVasya: look for the slang meaning of "to fap" 2011-10-25T13:42:43 UncleVasya: races/sport/card settings == Boring 2011-10-25T13:42:46 lets fact it 2011-10-25T13:42:49 *** master_ninja has joined #aichallenge 2011-10-25T13:42:56 *** choas has joined #aichallenge 2011-10-25T13:43:01 if we were good at sports we wouldn't be writing code 2011-10-25T13:43:16 we would be getting X million for kicking a ball around 2011-10-25T13:43:37 Also there is the simulation soccer league 2011-10-25T13:43:54 even a real robot one if you are interested 2011-10-25T13:44:43 raoul rojas, yea! 2011-10-25T13:49:29 *** cybsy has quit IRC (Ping timeout: 252 seconds) 2011-10-25T13:51:45 *** Redgis has quit IRC (Read error: Connection reset by peer) 2011-10-25T13:52:29 *** Redgis has joined #aichallenge 2011-10-25T13:52:50 *** Guest58853 has quit IRC (Quit: leaving) 2011-10-25T13:53:29 *** tdubellz has joined #aichallenge 2011-10-25T13:55:31 *** cybsy has joined #aichallenge 2011-10-25T13:55:54 *** clckwrk1 has joined #aichallenge 2011-10-25T13:56:00 *** lvqinghouer has quit IRC (Ping timeout: 240 seconds) 2011-10-25T13:57:06 *** Palmik has joined #aichallenge 2011-10-25T13:57:08 *** UncleVasya has quit IRC (Read error: Connection reset by peer) 2011-10-25T13:58:18 *** clckwrk has quit IRC (Ping timeout: 255 seconds) 2011-10-25T13:58:20 *** JamesMG has joined #aichallenge 2011-10-25T13:59:09 *** UncleVasya has joined #aichallenge 2011-10-25T13:59:34 *** JamesMG_ has quit IRC (Read error: Connection reset by peer) 2011-10-25T14:02:01 @later tell janzert I've got a color test I want you to look at http://imgur.com/wqxor 2011-10-25T14:02:01 McLeopold: Ready to serve, my lord. 2011-10-25T14:03:21 *** Racko has quit IRC (Ping timeout: 265 seconds) 2011-10-25T14:03:49 *** fap-machine has quit IRC (Quit: fap-machine) 2011-10-25T14:03:50 *** aaa has quit IRC (Ping timeout: 265 seconds) 2011-10-25T14:03:55 *** Puj has joined #aichallenge 2011-10-25T14:08:43 *** onensora has quit IRC (Ping timeout: 245 seconds) 2011-10-25T14:09:08 *** cirno_the_greate has joined #aichallenge 2011-10-25T14:09:10 *** ig has quit IRC (Read error: No route to host) 2011-10-25T14:09:15 jmcarthur: hallo 2011-10-25T14:09:39 you mentioned at forum about packaging haskell libraries with bot sources 2011-10-25T14:09:47 could you describe process 2011-10-25T14:10:08 which you referred as "alternative solution", please 2011-10-25T14:10:15 *** ig has joined #aichallenge 2011-10-25T14:11:46 *** onensora has joined #aichallenge 2011-10-25T14:13:19 McLeopold: how about putting the black borders back and only color in the # of ants graph? 2011-10-25T14:13:50 not the bar graphs? 2011-10-25T14:14:04 I don't think so 2011-10-25T14:14:21 it seems like too much brown 2011-10-25T14:14:23 I didn't mean to remove the borders, I think I just did the fill/stroke in the wrong order 2011-10-25T14:14:44 maybe with the border the other brown would look better 2011-10-25T14:17:10 *** jcdjcd has joined #aichallenge 2011-10-25T14:17:22 *** Relax has quit IRC (Ping timeout: 265 seconds) 2011-10-25T14:18:08 what's the correct sql test for not equal to null, != null ok? 2011-10-25T14:18:43 ahh maybe "is not null" 2011-10-25T14:19:03 indeed 2011-10-25T14:19:08 http://imgur.com/vY6Pr 2011-10-25T14:19:55 *** jcdny has quit IRC (Ping timeout: 252 seconds) 2011-10-25T14:20:46 *** thestinger has joined #aichallenge 2011-10-25T14:21:09 yeah, that looks nicer I think 2011-10-25T14:21:18 no, to many lines 2011-10-25T14:21:20 except the top two titles need to be black 2011-10-25T14:21:44 scores and # of ants with a black background? 2011-10-25T14:22:20 could try it I suppose, but the white lets it blend in with the surrounding background 2011-10-25T14:22:32 I'd try all three with a black background I think 2011-10-25T14:22:34 don't you think the black/white borders are a bit much right next to each other? 2011-10-25T14:23:08 I don't think so 2011-10-25T14:25:44 *** |UncleVasya| <|UncleVasya|!kvirc@178-133-209-62.dialup.umc.net.ua> has joined #aichallenge 2011-10-25T14:26:18 *** UncleVasya has quit IRC (Read error: Connection reset by peer) 2011-10-25T14:26:28 http://imgur.com/a6E2e 2011-10-25T14:27:02 *** xar0l has quit IRC (Ping timeout: 265 seconds) 2011-10-25T14:27:17 @seen jmcarthur 2011-10-25T14:27:17 cirno_the_greate: jmcarthur was last seen in #aichallenge 6 hours, 54 minutes, and 40 seconds ago: @later tell amstan sorry about lambdabot. i was demonstrating a couple things and intended to kick her out when we were done, but forgot 2011-10-25T14:27:31 hmm, so I guess that doesn't work so well 2011-10-25T14:28:35 can you go back to the last one and get rid of the white border inside the black around the line graph? 2011-10-25T14:29:14 there's no way to get debug info from the server, right? 2011-10-25T14:30:52 *** Shiiben has joined #aichallenge 2011-10-25T14:30:54 *** master_ninja has quit IRC (Quit: ..) 2011-10-25T14:31:14 also, the spec page says: "end" indicates that the game is over, the winner of the game will receive information for the final state of the game following this, should they wish to use it for local testing. 2011-10-25T14:35:16 *** mihanik has joined #aichallenge 2011-10-25T14:36:08 *** lvqinghouer has joined #aichallenge 2011-10-25T14:37:20 my bot doesn't output after the "end" command has given 2011-10-25T14:37:44 four minutes to next game with new version, i can't wait :| 2011-10-25T14:37:50 :-) 2011-10-25T14:37:53 good luck 2011-10-25T14:38:17 "no thanks" 2011-10-25T14:38:39 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-10-25T14:38:47 i wonder if in english it works the same way: if someone wishes you luck, you say "no thanks" instead of "thanks".. 2011-10-25T14:39:02 *** g0llum has joined #aichallenge 2011-10-25T14:39:24 in the starter bots, whats the point of keeping seen_water? 2011-10-25T14:40:00 Garf: you only get water information once, on the first turn you see it. 2011-10-25T14:40:07 olexs: yes, and? 2011-10-25T14:40:16 Garf: you don't know the map 2011-10-25T14:40:21 olexs: the bots keep a map[x][y] 2011-10-25T14:40:24 you need to collect information as you move 2011-10-25T14:41:13 *** cirno_the_greate has quit IRC (Remote host closed the connection) 2011-10-25T14:41:36 Garf: okay, I don't know my way around the starter bots. maybe there is a reason for the redundancy 2011-10-25T14:42:20 actually, the python start bot doesn't do this 2011-10-25T14:42:26 I guess the one i'm looking at is outdated 2011-10-25T14:42:30 *** DrSat has quit IRC (Ping timeout: 265 seconds) 2011-10-25T14:42:30 *** tester has quit IRC (Ping timeout: 265 seconds) 2011-10-25T14:42:41 *** irchs has joined #aichallenge 2011-10-25T14:44:57 duh, first game with some randombots, that was short :| 2011-10-25T14:45:04 Garf, the sooner, you learn, that those starter bots are just to show you the game io, the better, .. 2011-10-25T14:45:24 Fluxid: first games after reupload are always vs low-ranked bots 2011-10-25T14:45:27 i wish new version would be tested with participiants from around the same place as older versions 2011-10-25T14:45:34 olexs: i know, this sucks :( 2011-10-25T14:46:05 Fluxid: it's controversial, but there are reasons to reset the ranking with every upload 2011-10-25T14:46:33 next version must be very good to risk reupload 2011-10-25T14:46:45 *** |UncleVasya| <|UncleVasya|!kvirc@178-133-209-62.dialup.umc.net.ua> has quit IRC (Ping timeout: 244 seconds) 2011-10-25T14:46:56 I test extensively on TCP first, and only upload when I'm completely happy with the current version 2011-10-25T14:47:10 i did it too, and ested with older version.. 2011-10-25T14:47:28 but that's not all i have in mind, gonna spend more time on it 2011-10-25T14:48:07 *** xar0l has joined #aichallenge 2011-10-25T14:49:42 *** Larose has joined #aichallenge 2011-10-25T14:50:53 *** FLashM has joined #aichallenge 2011-10-25T14:51:45 I did it! I mean, they did it! http://ants.fluxid.pl/replay.5621 2011-10-25T14:52:28 *** Knekkebjoern has quit IRC (Read error: Connection reset by peer) 2011-10-25T14:53:14 GG rotfl 2011-10-25T14:53:15 http://aichallenge.org/visualizer.php?game=11487 what the hell 2011-10-25T14:53:43 FLashM: LOL 2011-10-25T14:54:01 awesome! 2011-10-25T14:54:17 i must do it! 2011-10-25T14:54:20 *** Knekkebjoern has joined #aichallenge 2011-10-25T14:54:21 *** clckwrk1 has quit IRC (Read error: Connection reset by peer) 2011-10-25T14:54:31 *** clckwrk has joined #aichallenge 2011-10-25T14:54:34 after razing hill i'll write "PWNED" near it! 2011-10-25T14:55:13 *** tncardoso has quit IRC (Quit: bye) 2011-10-25T14:55:22 *** ztfw has joined #aichallenge 2011-10-25T14:55:26 The hardest thing is to decide, when to form this label. I thought my conditions was a bit optimistic, but it turned out it could really happen 2011-10-25T14:55:36 as your hill is being razed 2011-10-25T14:57:08 No, you still can win the round even if your hill is razed, and if you trying to show something, you are baked for sure 2011-10-25T14:57:32 sure, doesnt invalidate my point :P 2011-10-25T14:58:08 Yup. =) 2011-10-25T14:58:58 lets convert the player in a "boom", "bang", "tjaw",-style :p 2011-10-25T14:59:42 *** skunx has joined #aichallenge 2011-10-25T14:59:47 * avdg will submit a game without moving ants afterwards 2011-10-25T15:00:04 *** Craklyn has joined #aichallenge 2011-10-25T15:00:11 is there any difference in the way I should treat "dead ant" and normal ground? 2011-10-25T15:00:34 You should sound the alarm if you see one =) 2011-10-25T15:00:44 *** msqrdstr has left #aichallenge ("i don't hate people. i just feel better when they aren't around. - bukowski") 2011-10-25T15:00:57 Draivin: Nothing that you need to do - unless you want to know if an ant recently died. 2011-10-25T15:01:32 thanks =) 2011-10-25T15:01:50 *** sur has quit IRC (Ping timeout: 265 seconds) 2011-10-25T15:03:07 now that I think about it, it sure is obvious, I was cloning my live ents list by the end of the turn, and removing from the list each ant I received at the beggining of the next, totally forgot about this then. 2011-10-25T15:03:14 ants* 2011-10-25T15:04:41 *** JamesMG_ has joined #aichallenge 2011-10-25T15:07:47 *** amstan has joined #aichallenge 2011-10-25T15:07:47 *** ChanServ sets mode: +o amstan 2011-10-25T15:07:52 *** JamesMG has quit IRC (Ping timeout: 258 seconds) 2011-10-25T15:08:18 *** FLashM has quit IRC (Quit: Page closed) 2011-10-25T15:12:57 *** lavalamp has quit IRC (Ping timeout: 265 seconds) 2011-10-25T15:14:46 *** analyst74 has joined #aichallenge 2011-10-25T15:18:50 *** Saulzar has quit IRC (Ping timeout: 276 seconds) 2011-10-25T15:19:45 *** herpderp has joined #aichallenge 2011-10-25T15:19:57 *** herpderp is now known as herpdiederpderp 2011-10-25T15:20:48 Hello, when working in a team while having a normal account....does it count as having two accounts? 2011-10-25T15:22:01 if you are submitting a personal bot and are on a team submitting a team bot, yes that is two accounts 2011-10-25T15:22:11 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-10-25T15:23:13 So that is prohibited? 2011-10-25T15:24:44 *** danielharan has quit IRC (Remote host closed the connection) 2011-10-25T15:25:27 yes 2011-10-25T15:26:02 Then is it possible to delete your account? 2011-10-25T15:27:56 *** chaudeau has quit IRC (Ping timeout: 264 seconds) 2011-10-25T15:28:37 *** Saulzar has joined #aichallenge 2011-10-25T15:31:01 *** Knekkebjoern1 has joined #aichallenge 2011-10-25T15:32:20 *** Knekkebjoern has quit IRC (Ping timeout: 260 seconds) 2011-10-25T15:33:20 *** retybok has joined #aichallenge 2011-10-25T15:33:33 *** Accoun has quit IRC () 2011-10-25T15:33:47 Are there alternative tcp servers? The current one is very slow, and it seems to play almost only maze maps... 2011-10-25T15:34:30 retybok: which one are you currently using? 2011-10-25T15:35:51 The one described in this post: http://aichallenge.org/forums/viewtopic.php?f=25&t=1505 2011-10-25T15:36:40 there is one on http://ants.fluxid.pl/ , if thats a valid one 2011-10-25T15:36:56 herpdiederpderp: not that i'm aware, at the moment. you should be fine if you simply abandon it 2011-10-25T15:37:03 *** dvladim has quit IRC (Ping timeout: 245 seconds) 2011-10-25T15:37:22 but there should be a way to delete it, so that we don't waste time giving it games if you uploaded a starter already 2011-10-25T15:37:32 avdg: it does seem faster, I think I'll give it a try. Thanks! 2011-10-25T15:39:20 Zannick: The problem is that I have uploaded 8 versions of the AI and now I want to work with some friends in a Team 2011-10-25T15:39:34 should I just re-upload the starter package to show that I have abandoned it? 2011-10-25T15:39:43 *** sunshowers has joined #aichallenge 2011-10-25T15:40:48 *** mcstar has left #aichallenge ("WeeChat 0.3.5") 2011-10-25T15:40:54 um, unless one of the site admins can delete your account 2011-10-25T15:41:33 at the very least, uploading the starter will keep you out of the top parts of the finals ;) 2011-10-25T15:41:55 or even a bot that immediately exits 2011-10-25T15:42:38 i'd ask janzert or amstan about that 2011-10-25T15:43:14 I'll ask them later when they are on IRC then :) 2011-10-25T15:44:59 *** Saulzar has quit IRC (Ping timeout: 252 seconds) 2011-10-25T15:45:18 Zannick: i can admin you so you can do it too 2011-10-25T15:48:31 *** icefox has quit IRC (Quit: icefox) 2011-10-25T15:49:05 *** icefox has joined #aichallenge 2011-10-25T15:50:54 amstan: excellent 2011-10-25T15:51:24 *** savaron has joined #aichallenge 2011-10-25T15:51:29 then i can be a real admin instead of just "that guy on irc who bans trolls" 2011-10-25T15:52:02 <_flag> Any one in charge of cutoff rules here? 2011-10-25T15:52:17 I think I can make a decision 2011-10-25T15:52:50 yes 2011-10-25T15:53:10 *** Accoun has joined #aichallenge 2011-10-25T15:53:21 <_flag> Do you think this game should have been cutoff? Obviously if I rushed his hill at the end I could have killed him, but I think the engine is assuming I didn't have to go through a bunch of tunnels to get to the hive: http://ants.fluxid.pl/replay.5665 2011-10-25T15:53:36 <_flag> And yes I know that's the tcp server and not the official one, but I'm assuming it's the same code 2011-10-25T15:54:05 *** Lacco has joined #aichallenge 2011-10-25T15:54:26 the maps and cutoff turn limits need to be synced 2011-10-25T15:54:26 *** clckwrk has quit IRC (Read error: Connection reset by peer) 2011-10-25T15:54:27 look at the message, then watch the code ;-) 2011-10-25T15:54:40 we haven't enforced max hill distance in the map generator yet 2011-10-25T15:55:13 <_flag> Okay, thanks 2011-10-25T15:55:14 so, the cutoff turn amount should be about the same as max hill distance 2011-10-25T15:55:34 <_flag> From watching it I assume it takes into account the number of ants my opponent has left, correct? 2011-10-25T15:55:35 (meh, I wonder myself what the code looks like for that specific part) 2011-10-25T15:55:42 which starts its count when you have 90% of the ants 2011-10-25T15:55:54 *** clckwrk has joined #aichallenge 2011-10-25T15:58:32 *** savaron has quit IRC (Quit: Page closed) 2011-10-25T15:59:20 I think that game is a legitimate cutoff though 2011-10-25T15:59:53 by the time you had 90% of the ants you should have been able to get too his hill too, right? 2011-10-25T16:00:38 <_flag> I definately could have won that, but my bot would have to recognize that I have an insurmountable lead and rush his hill 2011-10-25T16:01:05 <_flag> I'm not saying I can't do that, I'm just not sure if it should be a requirement for a bot not to be cutoff 2011-10-25T16:01:14 you locate his hill around turn 500 and have another 500 turns to mount some sort of offense though 2011-10-25T16:01:41 <_flag> Correct but I also want to gain control of the map so he doesn't take all the food 2011-10-25T16:02:10 <_flag> And due to the low rate of food spawn I only have full control over the map at around turn ~900 2011-10-25T16:02:28 <_flag> I'm trying to do the least risky play possible 2011-10-25T16:07:18 *** halligalli has joined #aichallenge 2011-10-25T16:11:20 you could have just killed his last remaining ant at the end and won as well :} 2011-10-25T16:11:40 <_flag> I was about to do that 2011-10-25T16:12:08 janzert: I wouldn't mind the cutoff percent being lowered and the turns being razed 2011-10-25T16:12:15 raised :) 2011-10-25T16:13:49 *** Lacco has quit IRC (Quit: Page closed) 2011-10-25T16:19:09 *** godloveshorses has joined #aichallenge 2011-10-25T16:21:08 *** JamesMG has joined #aichallenge 2011-10-25T16:23:36 *** sunshowers has quit IRC (Quit: Page closed) 2011-10-25T16:24:32 *** JamesMG_ has quit IRC (Ping timeout: 258 seconds) 2011-10-25T16:25:02 *** cleung24 has joined #aichallenge 2011-10-25T16:25:18 *** xathis_ has joined #aichallenge 2011-10-25T16:26:17 *** cleung24 is now known as Darrow 2011-10-25T16:28:00 *** xathis has quit IRC (Ping timeout: 240 seconds) 2011-10-25T16:29:46 well... that totally broke my crappy pathfinding 2011-10-25T16:29:50 http://ash.webfactional.com/replay.1313 2011-10-25T16:29:53 *** SharkMonkey has joined #aichallenge 2011-10-25T16:30:07 one of my hills (near the bottom right) is fully surrounded by water o_o 2011-10-25T16:31:24 This is what happens when you outsource construction labor 2011-10-25T16:31:26 buy (local) 2011-10-25T16:35:31 *** grumpy has joined #aichallenge 2011-10-25T16:36:48 *** Macuyiko has quit IRC (Ping timeout: 240 seconds) 2011-10-25T16:36:50 *** grumpy has quit IRC (Client Quit) 2011-10-25T16:36:55 *** Morley has joined #aichallenge 2011-10-25T16:37:12 *** delt0r_ has quit IRC (Ping timeout: 240 seconds) 2011-10-25T16:37:25 *** Morley has quit IRC (Client Quit) 2011-10-25T16:37:56 *** rascal999 has joined #aichallenge 2011-10-25T16:39:15 can the findings of individual ants be shared among all player ants? 2011-10-25T16:40:47 *** cirno_the_greate has joined #aichallenge 2011-10-25T16:40:54 @seen jmcarthur 2011-10-25T16:40:54 cirno_the_greate: jmcarthur was last seen in #aichallenge 9 hours, 8 minutes, and 17 seconds ago: @later tell amstan sorry about lambdabot. i was demonstrating a couple things and intended to kick her out when we were done, but forgot 2011-10-25T16:41:00 baww 2011-10-25T16:41:27 any haskell guys around? 2011-10-25T16:43:14 *** Jakinho has joined #aichallenge 2011-10-25T16:43:58 *** Jakinho has left #aichallenge 2011-10-25T16:44:28 *** Jakinho has joined #aichallenge 2011-10-25T16:44:35 nice, another map my pathfinding fails at :P 2011-10-25T16:44:41 this one is pretty funny though 2011-10-25T16:45:00 http://ompldr.org/vYXphaA 2011-10-25T16:45:05 rascal999: If you mean, can the findings of your individual ants can be seen by all of your ants, yes. 2011-10-25T16:45:25 rascal999: You control them all at once 2011-10-25T16:45:41 I pretend my hills aren't passable so I understand why that ant is stuck, but not why he went there in the first place :\ 2011-10-25T16:45:51 yay bugs 2011-10-25T16:46:21 Hah, my latest bot won its first game, by virtue of everyone else timing out. 2011-10-25T16:46:56 I would like to ask, if our team wants to change the name, because I begun individual and now we are a team, how can we do that? 2011-10-25T16:47:10 *** ztfw has quit IRC (Remote host closed the connection) 2011-10-25T16:47:17 *want 2011-10-25T16:47:42 *** guruuh has joined #aichallenge 2011-10-25T16:47:51 Jakinho: wants was correct. :) 2011-10-25T16:48:03 *** rascal999 has quit IRC (Ping timeout: 252 seconds) 2011-10-25T16:48:23 :D oh yes 2011-10-25T16:49:45 *** Palmik has quit IRC (Remote host closed the connection) 2011-10-25T16:49:45 *** tan-d42 has joined #aichallenge 2011-10-25T16:50:21 *** delt0r_ has joined #aichallenge 2011-10-25T16:51:00 hi 2011-10-25T16:51:54 *** aidan1 has joined #aichallenge 2011-10-25T16:53:00 *** Shiiben has quit IRC (Ping timeout: 265 seconds) 2011-10-25T16:53:15 *** ltriant has joined #aichallenge 2011-10-25T16:54:17 it would be nice to have graphs, for example the score for each ai version 2011-10-25T16:54:37 *** clckwrk1 has joined #aichallenge 2011-10-25T16:54:37 *** clckwrk has quit IRC (Read error: Connection reset by peer) 2011-10-25T16:55:22 Admins, please, is there any possibility? 2011-10-25T16:55:46 *** clckwrk1 is now known as clckwrk 2011-10-25T16:57:09 Would it be rude to highlight the staff here? If we have a question like Jakinho's, which might only be answerable by them? 2011-10-25T16:59:21 xathis_: hi, you're from lübeck too right? 2011-10-25T16:59:32 Jakinho: what's the current name and new name you'd like? 2011-10-25T16:59:51 *** herpdiederpderp has quit IRC (Quit: Page closed) 2011-10-25T17:01:10 Current name: Jakinho, new name: Kudlaci 2011-10-25T17:01:46 *** clckwrk1 has joined #aichallenge 2011-10-25T17:02:30 jix_: i am 2011-10-25T17:02:33 *** clckwrk1 has left #aichallenge 2011-10-25T17:03:06 jix_: you're stuying computer science as well i guess? 2011-10-25T17:03:12 Jakinho: should be set 2011-10-25T17:04:02 OK, thanks, and can we please change password too? 2011-10-25T17:05:20 Hey everyone, what python libraries are available for the aichallenge? I hear numpy/scipy are, anything else? 2011-10-25T17:05:24 *Home 2011-10-25T17:05:30 *** clckwrk has quit IRC (Ping timeout: 255 seconds) 2011-10-25T17:05:30 *** rabuf` has joined #aichallenge 2011-10-25T17:05:35 *by default, not with anything I include 2011-10-25T17:06:31 my program works fine in local with the tools, but when I use it on aichallenge.org or ants.fluxid.pl, it always timeouts; what could cause this kind of problem ? 2011-10-25T17:06:40 *** rabuf has quit IRC (Read error: Operation timed out) 2011-10-25T17:07:34 Jakinho: I'm not sure on how to do the password reset offhand, if you see amstan around he can probably do it for you 2011-10-25T17:07:51 hopefully we'll have an automated system for it soon 2011-10-25T17:08:31 xathis_: yeah 2011-10-25T17:09:38 @amstan: Can our team change password, please? 2011-10-25T17:09:39 Jakinho: Run as fast as you can and don't look back. 2011-10-25T17:10:34 contestbot: I don't understand. 2011-10-25T17:10:35 Jakinho: User error, it's not my fault. 2011-10-25T17:11:17 *** skunx has quit IRC (Quit: leaving) 2011-10-25T17:11:34 off topic: singup list is missing few countries 2011-10-25T17:13:12 *** Diede has joined #aichallenge 2011-10-25T17:14:04 How are new affiliations created? 2011-10-25T17:15:13 *** tncardoso has joined #aichallenge 2011-10-25T17:15:27 *** dauryg has quit IRC (Quit: Page closed) 2011-10-25T17:15:31 I wrote my bot in Javascript and I ran different tests on 8p multi hill mazes and 5000 turns, everything works fine. But online my bot crashes each and every game, cause a bad skill factor, it is 27 now and the game it is based on was completely dominated by me, score and ants wise. I killed 3 of the 4 others and then my bot crashed. 2011-10-25T17:15:34 Can the color of your own bot be changed for replays? If you view a replay with user=id in the url your own bot is rather hard to see. 2011-10-25T17:15:53 *** dominater has joined #aichallenge 2011-10-25T17:15:54 Any idea why it is crashing, and how do I solve it? I don't even get an error message 2011-10-25T17:19:38 *** ardillaroja has joined #aichallenge 2011-10-25T17:23:34 *** just_visiting has joined #aichallenge 2011-10-25T17:24:13 *** retybok has quit IRC (Quit: leaving) 2011-10-25T17:24:37 *** just_visiting has left #aichallenge 2011-10-25T17:24:56 Diede: did you run it locally? 2011-10-25T17:25:19 Yes, on different 8p multi hill mazes with 5000 turns 2011-10-25T17:25:23 *** dominater has quit IRC (Ping timeout: 265 seconds) 2011-10-25T17:25:32 wait, I'll search my script for the options 2011-10-25T17:26:04 ok 2011-10-25T17:26:11 if you do ./playgame.py you get the options as well 2011-10-25T17:26:36 *** ajhager has quit IRC (Quit: ajhager) 2011-10-25T17:26:42 use flags —log_dir 2011-10-25T17:26:51 and -E 2011-10-25T17:26:57 *** povik has joined #aichallenge 2011-10-25T17:27:00 hi 2011-10-25T17:27:28 Diede: if you use these flags, you can find the log files in the specified location 2011-10-25T17:27:51 (I'm using flags -IOE for logging) 2011-10-25T17:28:22 avdg: Thanks, I'll give that a run 2011-10-25T17:28:48 avdg: But logging isn't gonna make my bots crash locally, is it ? They are crashing online for no known reason 2011-10-25T17:29:15 *** ardillaroja has quit IRC (Ping timeout: 265 seconds) 2011-10-25T17:29:29 loggin just gives you a reason why it might have crashed ;-) 2011-10-25T17:29:38 *logging 2011-10-25T17:29:53 Diede, You might just have to keep running simulations until it happens locally :) 2011-10-25T17:30:24 Or look at the state when the bot crashed online, and reason out what might have happened. 2011-10-25T17:30:46 Am I allowed to write debug info from the bot to a server ? (I guess not, but I wanna try) 2011-10-25T17:30:53 and if you want to have debug messages, just use process.stderr.write ;-) 2011-10-25T17:30:57 Diede, My bot used to crash only when playing weaker opponents. I had a ton of ants. 2011-10-25T17:31:09 Yeah, I've been looking at the states, I see no simularities 2011-10-25T17:31:23 Yes, I also have a ton of ants 2011-10-25T17:31:47 When I played locally I was playing against variations of my own bots that didn't let me get so strong. But online, against weaker oponents I'd crash. 2011-10-25T17:32:11 *** kevlar has joined #aichallenge 2011-10-25T17:32:16 Only found the bug locally when playing against the example bots. And that only happened like 1 in 5 times 2011-10-25T17:32:37 *** kevlar is now known as Guest97303 2011-10-25T17:32:37 well, maybe a manual cap can help (just don't move the ant on top of the hill if you reached the limit) 2011-10-25T17:32:40 Why does it crash ? Can I request to put me up against harder oponents ? 2011-10-25T17:33:00 tcp server has plenty tough opponents 2011-10-25T17:33:01 Diede: nope you can't choose your opponents 2011-10-25T17:33:11 most of them are tough, I'd say 2011-10-25T17:33:19 but to log crashes, you have to run it locally I'm afraid 2011-10-25T17:33:25 or use a tcp game 2011-10-25T17:33:30 *** Jakinho has quit IRC (Quit: Page closed) 2011-10-25T17:33:44 like http://ants.fluxid.pl/ 2011-10-25T17:33:55 (there are more links on the forum) 2011-10-25T17:34:29 I'll run some variated scenarios locally with logging, thanks for the help guys. Good night :) 2011-10-25T17:35:00 avdg: Thanks for the link, I'll look into that 2011-10-25T17:37:09 "And remember, testing is the future" :-) 2011-10-25T17:37:58 *** tobias has joined #aichallenge 2011-10-25T17:38:24 *** tobias is now known as Guest70576 2011-10-25T17:39:31 hi everybody! i found the ai challenge yesterday and plan to participate by writing in python. my question is if it is allowed to change the ants class in ants.py, or if that is given by the server in the end? 2011-10-25T17:40:16 Guest70576, you can change it 2011-10-25T17:40:52 alright, thanks 2011-10-25T17:40:52 well, depends on which ants.py 2011-10-25T17:41:09 Guest70576, the server just communicates through stdin/stdout to your bot 2011-10-25T17:41:12 (there is a server side script called ants.py as well :p) 2011-10-25T17:41:42 Guest70576, so you can change it as long as you keep reading the correct data in and outputing the correct data :P 2011-10-25T17:41:49 avdg, Be careful confusing new comers :) 2011-10-25T17:41:51 well i mean the one in the starter package, where it stores your ants and stuff ;) 2011-10-25T17:42:00 so i guess it's alright ;) 2011-10-25T17:42:05 SharkMonkey: Yeah, I was thinking too late :/ 2011-10-25T17:42:32 avdg, Guest70576 seems like he has a handle on things, anyway :) 2011-10-25T17:43:32 SharkMonkey, so you can confuse newcomers, you just have to do it carefully? ;) 2011-10-25T17:43:35 *** Guest97303 has quit IRC (Quit: Page closed) 2011-10-25T17:45:44 Diede: hmm, so your the only one my country doing js as well right? (fun fact) 2011-10-25T17:46:25 *** guruuh has quit IRC (Quit: Page closed) 2011-10-25T17:47:00 avdg:Hah cool, I didn't know you where from Belgium :-) Yeah, I usually work in C# but I couldn't get that to work 2011-10-25T17:47:13 And I've been doing some stuff in node lately so yeah 2011-10-25T17:47:20 *** spindoctor has joined #aichallenge 2011-10-25T17:47:54 bweh, js isn't the easiest language either, but it at least works for now :-) 2011-10-25T17:48:15 *** irchs has quit IRC (Quit: irchs) 2011-10-25T17:49:10 Alexer, Exactly :P 2011-10-25T17:50:08 *** Diede has quit IRC (Quit: Page closed) 2011-10-25T17:51:46 *** ksteele has joined #aichallenge 2011-10-25T17:52:18 *** rufus_ has joined #aichallenge 2011-10-25T17:52:24 *** halligalli has quit IRC (Quit: Page closed) 2011-10-25T17:55:18 *** Rinum has joined #aichallenge 2011-10-25T17:56:45 In the contesnt, do people tend to hold their bots back when it comes near to send your final submissions? 2011-10-25T17:57:02 They don't want to reveal their strategies and have someone make counters to them 2011-10-25T17:57:11 *In past contests 2011-10-25T17:57:50 Like ebay sniping. People don't place their highest bids until the last minute. 2011-10-25T17:58:11 you could risk with a buggy bot 2011-10-25T17:58:24 i don't think that happens a lot because unless you submit or use the tcp servers you don't know how well it actually performs 2011-10-25T17:58:41 SharkMonkey: the more and earlier you test, the better ;-) 2011-10-25T17:58:46 *** choas has quit IRC (Read error: Operation timed out) 2011-10-25T17:59:16 especially if you don't run under the time limit on the contest workers but you do on your own machine(s). 2011-10-25T18:00:18 I'm developing on a netbook, so there is little risk of that :) 2011-10-25T18:00:42 *** Guest70576 has quit IRC (Quit: Page closed) 2011-10-25T18:00:49 oh, my local time limit is 300ms (500 on server) 2011-10-25T18:00:50 But yeah. There are a lot of unkowns when playing on the server and its hard to find out what works without trying 2011-10-25T18:01:02 that should help against most timeouts 2011-10-25T18:01:30 how are the bots in the final competition decided? i don't suppose everyone will get in 2011-10-25T18:01:39 cyphase, Why not? 2011-10-25T18:01:50 cyphase: they run a final contest match 2011-10-25T18:01:56 same ranking system we currently have 2011-10-25T18:02:07 *** Rinum has quit IRC (Ping timeout: 265 seconds) 2011-10-25T18:02:27 it just seems from what i read that the amount of games run for the final competition would make it prohibitive for 2000+ bots 2011-10-25T18:02:33 maybe i got the wrong impression 2011-10-25T18:03:44 assuming 20 player/min, each player playing 10 games (20000), that's 1000 minutes 2011-10-25T18:03:45 Well they don't have to play every single bot against every other bot 2011-10-25T18:04:15 with their matching algorithm, your ranking stabilizes after just a few games 2011-10-25T18:04:17 1000minutes is less than a day 2011-10-25T18:04:39 I don't think the servers are hitting 20 players/min 2011-10-25T18:04:49 unless there are more workers 2011-10-25T18:04:59 I believe they have funding for some extra 2011-10-25T18:05:55 shark: no.. most people put their real final bot up, long before the deadline 2011-10-25T18:06:11 (hmm, unless the average players is higher than 6?) 2011-10-25T18:06:20 *** aidan1 has quit IRC (Quit: WeeChat 0.3.5) 2011-10-25T18:06:22 *average players each game 2011-10-25T18:08:22 *** onensora has quit IRC (Ping timeout: 260 seconds) 2011-10-25T18:08:23 I don't really know the details of plan, but if they decide to run the final round for a week, that wouldn't cause riots 2011-10-25T18:08:38 the final contest will run for about a week 2011-10-25T18:08:41 at least planet wars did 2011-10-25T18:09:01 planetwars was roughly 4days I think 2011-10-25T18:12:01 *** SharkMonkey has quit IRC (Ping timeout: 252 seconds) 2011-10-25T18:13:28 *** Islacrusez has joined #aichallenge 2011-10-25T18:14:00 *** analyst74 has quit IRC (Ping timeout: 240 seconds) 2011-10-25T18:15:06 *** JamesMG_ has joined #aichallenge 2011-10-25T18:15:54 *** xathis_ has quit IRC () 2011-10-25T18:16:42 *** olexs has quit IRC (Quit: Leaving.) 2011-10-25T18:17:52 *** Rinum has joined #aichallenge 2011-10-25T18:18:14 how much time does our code get per turn? 2011-10-25T18:18:38 Rinum, it's specified by turntime 2011-10-25T18:18:41 *** JamesMG has quit IRC (Ping timeout: 260 seconds) 2011-10-25T18:18:46 i believe it's 500ms currently on the servers 2011-10-25T18:19:05 ayo 2011-10-25T18:19:15 cool, thanks 2011-10-25T18:19:23 anyone present that is proficient with python2? 2011-10-25T18:19:36 Islacrusez, what's up? 2011-10-25T18:20:23 I'm trying to go through the tutorial, but the second page (about finding food) isn't very specific about where to insert/replace code 2011-10-25T18:20:37 let's see.. 2011-10-25T18:20:40 I've tried it a few ways and it either crashes or doesn't move 2011-10-25T18:21:34 ( http://aichallenge.org/ants_tutorial_step_2.php ) 2011-10-25T18:21:37 Islacrusez, the first chunk of code says "Create the following function after the do_move_direction function" 2011-10-25T18:21:57 Islacrusez, and the second chunk says "Now replace the default move with this:" 2011-10-25T18:22:08 yeah 2011-10-25T18:22:12 the first bit I think I managed 2011-10-25T18:22:54 *** Rinum has quit IRC (Quit: Page closed) 2011-10-25T18:24:18 Islacrusez, still not working? 2011-10-25T18:25:05 well I inserted the first bit of code and it didn't crash 2011-10-25T18:25:13 now lemme try the second bit, since that sounds really vague 2011-10-25T18:26:03 the default move you're replacing should have "# default move" above it 2011-10-25T18:27:21 http://paste.aichallenge.org/jEZT2/ :D 2011-10-25T18:27:26 that entire chunk of code between # default move and the whitespace? 2011-10-25T18:28:17 Islacrusez: pastebin your code 2011-10-25T18:28:26 Islacrusez, yea, that sounds right 2011-10-25T18:28:33 that loop 2011-10-25T18:28:36 I finally updated my code to have defense, still getting rocked by locutus 2011-10-25T18:28:43 I guess I need offense 2011-10-25T18:28:57 ant your posts! fiight! 2011-10-25T18:29:15 ok, did that and it didn't crash but it did nothing at all 2011-10-25T18:29:21 so I must have the first bit wrong? 2011-10-25T18:31:31 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-10-25T18:32:16 Islacrusez: if you pastebin your code, I can help http://pastebin.com/ 2011-10-25T18:32:52 McLeopold: ...flying ants? 2011-10-25T18:33:31 http://pastebin.com/W4SgqynT 2011-10-25T18:33:55 *** pvarga has joined #aichallenge 2011-10-25T18:34:03 oh, i see. "jesusbot" 2011-10-25T18:34:09 there you go 2011-10-25T18:34:50 Islacrusez: yep, there are some problems there :) 2011-10-25T18:35:19 ye don't say :P 2011-10-25T18:35:25 *** ikaros has joined #aichallenge 2011-10-25T18:35:38 I'll revert back to the last code that ran 2011-10-25T18:36:06 http://pastebin.com/YUm43aeG 2011-10-25T18:36:19 Islacrusez: you just need to indent the do_move_location 2011-10-25T18:36:29 what you did was end the function above 2011-10-25T18:37:30 but that's the indentation the tutorial gives you :S 2011-10-25T18:37:49 Why would turning psyco off decrease the occurence of unpredictable timeouts? 2011-10-25T18:38:31 do *you* know how psyco works? 2011-10-25T18:38:35 No 2011-10-25T18:38:46 Does it precompile to c? 2011-10-25T18:38:49 then if you turn it off, all timeouts become more predictable :P 2011-10-25T18:39:01 Good point, +t 2011-10-25T18:39:06 Where t =1 ? 2011-10-25T18:39:27 lol 2011-10-25T18:39:47 i don't know how to use psyco, so i can't help you with it 2011-10-25T18:40:06 whenever i've used psyco, it's been import psyco; psyco.full() 2011-10-25T18:40:24 *** amstan has quit IRC (Ping timeout: 240 seconds) 2011-10-25T18:40:35 Right, it's included in the startup package cyphase. 2011-10-25T18:40:40 For python that is. 2011-10-25T18:40:45 *** Seth_ has joined #aichallenge 2011-10-25T18:40:55 yea 2011-10-25T18:40:59 *** Seth_ is now known as seth1010 2011-10-25T18:41:26 Islacrusez: indentation is fixed now 2011-10-25T18:41:48 Zannick: since we are using python 2.7 on the server, I think pysco is out? 2011-10-25T18:42:22 Yup McLeopold 2011-10-25T18:42:59 McLeopold: and I thought I was just failing at interpreting instructions xD 2011-10-25T18:44:09 McLeopold: like i said, i wouldn't know :P 2011-10-25T18:44:17 i have only heard of psyco in this channel 2011-10-25T18:44:26 i use pure python myself when coding :P 2011-10-25T18:47:50 *** jesionaj has joined #aichallenge 2011-10-25T18:50:21 *** onensora has joined #aichallenge 2011-10-25T18:52:28 *** fr3tles5 has joined #aichallenge 2011-10-25T18:54:14 *** amstan has joined #aichallenge 2011-10-25T18:54:14 *** ChanServ sets mode: +o amstan 2011-10-25T18:58:35 *** Sunshowers has joined #aichallenge 2011-10-25T18:59:01 McLeopold: looks like the indentation is broken on 3 as well; looks like some unfortunate soul has to look through the whole lot? :S 2011-10-25T18:59:11 *** bansheechrome has joined #aichallenge 2011-10-25T18:59:35 (I've fixed my own code, at least) 2011-10-25T19:00:14 *** bansheechrome has quit IRC (Client Quit) 2011-10-25T19:02:43 *** Transformer has joined #aichallenge 2011-10-25T19:03:37 *** cybsy has quit IRC (Ping timeout: 258 seconds) 2011-10-25T19:05:58 *** Transformer has quit IRC (Excess Flood) 2011-10-25T19:07:26 *** levis501 has joined #aichallenge 2011-10-25T19:08:10 *** onensora has quit IRC () 2011-10-25T19:10:57 *** levis501 has joined #aichallenge 2011-10-25T19:11:43 *** shifu has quit IRC (Quit: Lost terminal) 2011-10-25T19:11:48 *** Redgis has quit IRC (Ping timeout: 244 seconds) 2011-10-25T19:11:52 *** Harpyon has quit IRC (Quit: Textual IRC Client: http://www.textualapp.com/) 2011-10-25T19:15:19 *** shark_teef is now known as modafinil 2011-10-25T19:16:08 *** spindoctor has quit IRC (Quit: Page closed) 2011-10-25T19:17:17 *** yoden has joined #aichallenge 2011-10-25T19:18:43 *** Mikustriker has joined #aichallenge 2011-10-25T19:19:20 looks like the ratings might be settling 2011-10-25T19:20:16 *** lvqinghouer has quit IRC (Quit: 离开) 2011-10-25T19:24:42 *** duckinator has left #aichallenge ("Narratore di falsità, i pantaloni hanno bruciata!") 2011-10-25T19:25:44 *** Mikustriker has quit IRC (Ping timeout: 265 seconds) 2011-10-25T19:27:07 what's a good way to see the output when you're bot crashes? 2011-10-25T19:27:36 i'm not sure why mine is crashing, as i've pasted input and it worked paste turn 1 2011-10-25T19:27:50 Are you running the game locally? 2011-10-25T19:28:08 yea 2011-10-25T19:28:23 linux, and it's in python 2011-10-25T19:28:41 i saw a command line earlier.. 2011-10-25T19:28:55 one second, looking up the option 2011-10-25T19:29:27 I guess I don't have it in my command history. I think did something like -O -log_output 2011-10-25T19:29:30 --log_output 2011-10-25T19:29:41 You can try searching for that in the .py scripts in the "tutorial" folder, I think? :P 2011-10-25T19:30:06 i have MyBot.sh and ( time python3 -mcProfile -o ./lol.profile ./src/MyBot.py3 ) 2> ./log.txt in it 2011-10-25T19:30:45 do you use the playgame.py? 2011-10-25T19:31:02 you can use -e flag to redirect to stderr i think 2011-10-25T19:31:11 cyphase, 2011-10-25T19:31:29 yea, i am 2011-10-25T19:32:07 trying it now 2011-10-25T19:37:53 *** Areks|2 has quit IRC (Ping timeout: 245 seconds) 2011-10-25T19:38:05 someone told me to add -eE to mine 2011-10-25T19:38:16 *** mleise has quit IRC (Quit: Leaving.) 2011-10-25T19:38:37 *** Israfel has quit IRC (Read error: Connection reset by peer) 2011-10-25T19:38:51 but that was python 2 under osx running from a script via terminal 2011-10-25T19:39:04 *** povik has left #aichallenge 2011-10-25T19:41:53 *** ksteele has quit IRC (Quit: Page closed) 2011-10-25T19:43:18 Fluxid: iss there a way to purge bots from the tcp sesrver? 2011-10-25T19:45:39 f4hy: what do you mean? 2011-10-25T19:46:23 Fluxid: I have played with many names now BOTNAMEv1 BOTNAMEv2 etc 2011-10-25T19:46:45 I realized one of them was trying to play using a broken build and timedout a bunch 2011-10-25T19:47:10 f4hy: it's sqllite, maybe i can purge some... but is this really a problem? 2011-10-25T19:47:12 just wondering if there iss support from removing it from the rankingss 2011-10-25T19:47:16 it isnt really 2011-10-25T19:47:22 no idea :) 2011-10-25T19:47:25 lemme check 2011-10-25T19:47:42 but if there was an easy way (like running the pythonscript with a flag or something) to remove it then eassy 2011-10-25T19:47:47 if not np 2011-10-25T19:47:49 *** Israfel has joined #aichallenge 2011-10-25T19:48:58 *** iris1 has joined #aichallenge 2011-10-25T19:49:56 Hey can someone versed in python buffers help me. 2011-10-25T19:50:06 I've now changed my do_turn method to only 'pass' 2011-10-25T19:50:13 *** Relax has joined #aichallenge 2011-10-25T19:50:14 And I still get unpredictable timeouts 2011-10-25T19:50:19 i'd need to write sql... i don't want to play with it now 2011-10-25T19:50:21 On different maps as well. 2011-10-25T19:50:45 Puj: did you profile it? 2011-10-25T19:51:42 Fluxid: thanks for looking. happy anting 2011-10-25T19:52:02 I tried getting the (time python -mcProfile ...) to work, but I don't know if cygwin hates me or if that syntax works in python3 or if it's the starter packages. But I did profile using runctx in my code on my main methods in do_turn, and that showed up clean. 2011-10-25T19:52:04 *** bumbleHEDGE has joined #aichallenge 2011-10-25T19:52:15 I am sad so many ants seem to have gotten really smart over the past few days 2011-10-25T19:52:17 And now I am literally doing nothing and it still times out, but like on turn 22 2011-10-25T19:52:45 yesterday I could do well, now they have attack code. 2011-10-25T19:54:03 well a sample bot will still hold the line above 1300 2011-10-25T19:54:06 xD 2011-10-25T19:54:08 Any idea Fluxid? 2011-10-25T19:55:19 *** Israfel_ has joined #aichallenge 2011-10-25T19:55:30 *** Israfel has quit IRC (Read error: Connection reset by peer) 2011-10-25T19:55:41 *** deltree__ has quit IRC (Quit: If you don't like it, don't suck at it) 2011-10-25T19:56:49 omg, why is this so hard 2011-10-25T19:56:58 *** Israfel_ has quit IRC (Client Quit) 2011-10-25T19:57:00 my bot is crashing at turn 1, and i cant figure out why 2011-10-25T19:57:09 *** Relax has quit IRC (Ping timeout: 265 seconds) 2011-10-25T19:57:20 *** Israfel has joined #aichallenge 2011-10-25T19:57:59 hello, I just created an account and I'm having trouble signing in. Do I ask for help here? 2011-10-25T19:58:36 have you activated the account via the confirmation email sent to the email address you provided? 2011-10-25T19:58:42 yes 2011-10-25T19:59:23 In that case I have no idea; but someone else in here might be able to help 2011-10-25T19:59:27 *** tassadar has joined #aichallenge 2011-10-25T19:59:30 Is it a password issue? Browser? 2011-10-25T19:59:30 *** staalebk has joined #aichallenge 2011-10-25T19:59:39 hey everyone, is there a way to check runtime of official games? 2011-10-25T20:00:52 I'm not sure, when I try to log in it thinks I spelled my username or password incorrectly but I've typed it correctly and even a couple variations in case I typed it wrong twice is there a way to confirm my password? 2011-10-25T20:01:14 also when I try to search for my name in the user search it doesn't find my name, is that normal? 2011-10-25T20:01:19 Send the password/reset it. 2011-10-25T20:01:39 how do I do that? 2011-10-25T20:02:36 hmm, don't see the option. Try making another account. 2011-10-25T20:03:01 I did, but it says I'm only allowed to have one and that I'll be disqualified if I try to make another 2011-10-25T20:03:17 Hmm, my bot timed out, while it runs fast on my computer... are there any known problems with the java compiler in use? 2011-10-25T20:03:22 aaah 2011-10-25T20:03:56 i was assuming there was a blank line between the data for every turn, as in the spec 2011-10-25T20:04:10 *** Relax has joined #aichallenge 2011-10-25T20:04:38 is there an email address I can send my problem to so someone can fix it? I can't find one on the website 2011-10-25T20:04:41 staalebk: lower the timelimit on your computer? 2011-10-25T20:04:55 avdg: how? 2011-10-25T20:05:19 and what is the limit on teh official servers? 2011-10-25T20:05:24 --turntime argument to playgame.py, do ./playgame.py -h 2011-10-25T20:05:27 by running ./playgame.py (which generates a help message) 2011-10-25T20:05:30 yep, that was it. the one line where i skip over the expected empty line 2011-10-25T20:05:39 ah, didn't even think of looking at that help message 2011-10-25T20:05:44 http://aichallenge.org/game_settings.php 2011-10-25T20:05:44 thanks 2011-10-25T20:06:02 *** swamp has joined #aichallenge 2011-10-25T20:06:27 wow, is turntime really .5 secs on official server? kinda low 2011-10-25T20:07:59 0.5 secs should still be plenty of time for my bot.. it goes trough 200 turns in not many seconds. 2011-10-25T20:08:06 But it still timed out on the very first move 2011-10-25T20:08:35 well, the default config in playgame is 1000 ms, so was assuming that 2011-10-25T20:08:56 tassadar: gpm would go half if turn time was bigger 2011-10-25T20:09:03 even with 50ms i still manage around 60 turns before I get enough ants to timeout on my computer... 2011-10-25T20:09:22 so I'm guessing something is taking way more time with the java compiler they use 2011-10-25T20:09:50 well, when i submitted first vers, i got timeout first turn too. 2011-10-25T20:09:59 then realized i was doing way too much precalculation 2011-10-25T20:09:59 staalebk: http://aichallenge.org/starter_packages.php 2011-10-25T20:10:17 it says the openjdk version there 2011-10-25T20:10:40 ah, maybe test with that version and see if I can reproduce the slowness 2011-10-25T20:10:52 idk why they don't just use standard sun java :P 2011-10-25T20:11:10 *** godloveshorses has quit IRC (Ping timeout: 265 seconds) 2011-10-25T20:11:12 tassadar: maybe because it's oracle java now ;) 2011-10-25T20:11:32 oracle java can't be redistributed iirc 2011-10-25T20:11:48 so linux distros can't package it and put it in their repos anymore 2011-10-25T20:12:02 *** doricatron has joined #aichallenge 2011-10-25T20:12:20 openJDK is basically the same speed as sun jdk tho 2011-10-25T20:13:01 does anyone knows what is the actual ants.py with which they run the bot for ranking? 2011-10-25T20:13:11 starterkit and tools folder have different 2011-10-25T20:13:18 *** Knekkebjoern1 has quit IRC (Quit: Leaving.) 2011-10-25T20:13:51 *** bumbleHEDGE has quit IRC (Quit: Page closed) 2011-10-25T20:14:12 doricatron: starter kit doesn't come with an ants.py 2011-10-25T20:14:56 wait, nvm 2011-10-25T20:15:01 unless you mean the python starter package 2011-10-25T20:15:07 the ants.py from there is different 2011-10-25T20:15:49 yeah python starter package 2011-10-25T20:16:15 for e.g. direction function is present in python starter package but is not present in tools/ants.py 2011-10-25T20:16:39 doricatron: one of the controls the engine, the other controls the one bot 2011-10-25T20:16:42 any admins here know worker server cpu specs? 2011-10-25T20:16:46 they're just named the same by coincidence 2011-10-25T20:16:59 tassadar: they're fairly fast, not sure offhand 2011-10-25T20:17:16 tassadar: if you're timing out, it's probably because of your algo 2011-10-25T20:17:32 @amstan: nah, just curious 2011-10-25T20:17:33 tassadar: I'm sorry Dave, err tassadar; I cannot 'amstan:'. 2011-10-25T20:18:07 can bots be multicore? 2011-10-25T20:18:12 or is server going to limit me to one? 2011-10-25T20:18:18 no 2011-10-25T20:18:19 yoden: they can't 2011-10-25T20:18:19 yoden: pretty sure rules say only 1 thread 2011-10-25T20:18:21 single thread 2011-10-25T20:18:25 lame 2011-10-25T20:18:26 :( 2011-10-25T20:18:27 lol 2011-10-25T20:18:39 What's the best way to play a .replay file? 2011-10-25T20:18:55 would be fun to mkae 80 threads and steal cpu time from others tho 2011-10-25T20:19:14 it's 2011, seems like any contest should include multi-core :P 2011-10-25T20:19:21 my bot is 10 lines. it connects to the cluster in my basement 2011-10-25T20:20:10 aichallenge: McLeopold epsilon * r5b7b2dd / (4 files): visualizer graph with dirt color background - http://git.io/HV57eg 2011-10-25T20:20:11 aichallenge: McLeopold epsilon * rb6a3fc6 / (ants/dist/starter_bots/java/Tile.java website/index.php): Merge branch 'epsilon' of github.com:aichallenge/aichallenge into epsilon - http://git.io/hbA9Kw 2011-10-25T20:20:13 multicore has tons of security and timing difficultys 2011-10-25T20:20:29 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-10-25T20:21:06 *** pvarga_ has joined #aichallenge 2011-10-25T20:22:39 are you explicitly notified when one of your ants dies? 2011-10-25T20:22:51 *** ig has quit IRC (Read error: No route to host) 2011-10-25T20:23:05 or does it just disappear? 2011-10-25T20:23:29 *** ig has joined #aichallenge 2011-10-25T20:23:37 @later tell janzert http://paste.aichallenge.org/Zd6pI/ ant paste is running with new colors, please take a look 2011-10-25T20:23:37 McLeopold: I come to serve. 2011-10-25T20:23:56 *** mleise has joined #aichallenge 2011-10-25T20:23:57 @later tell ChrisH_ look at colors on ant paste please http://paste.aichallenge.org/Zd6pI/ 2011-10-25T20:23:57 McLeopold: Ready to serve. 2011-10-25T20:24:00 *** pvarga has quit IRC (Ping timeout: 240 seconds) 2011-10-25T20:24:00 *** pvarga_ is now known as pvarga 2011-10-25T20:24:47 *** replore has joined #aichallenge 2011-10-25T20:25:29 Zug zug. Daboo. 2011-10-25T20:25:57 Puj: wut iz it? 2011-10-25T20:26:05 +1 amstan 2011-10-25T20:26:48 +2 for "Why do you keep touching me?" 2011-10-25T20:26:59 *** yoden has quit IRC (Read error: Connection reset by peer) 2011-10-25T20:27:03 that tickles 2011-10-25T20:27:20 You are a winner amstan 2011-10-25T20:27:33 i set all the messages for contestbot 2011-10-25T20:27:43 You are a cheater amstan 2011-10-25T20:27:52 oh, right 2011-10-25T20:28:02 so score/rank, is based purely on what place you get in each battle? 2011-10-25T20:28:11 Puj: it is not a good day to die? 2011-10-25T20:28:24 eg. winning 80:10, or 45:41 are both equal? 2011-10-25T20:28:35 Is -that- what the troll says? 2011-10-25T20:29:07 hmm, after all optimization, it seems that input is taking ~60 ms for later turns 2011-10-25T20:29:13 any suggestions for c++? 2011-10-25T20:29:17 using scanf right now 2011-10-25T20:29:44 if performance matters that much to you: write custom code 2011-10-25T20:30:01 *** grom358 has joined #aichallenge 2011-10-25T20:30:16 if two enemy ants meet in the middle of nowhere and kill each other, what do the bots see as input that turn? 2011-10-25T20:30:20 avdg: Found out the problem, It was the first time the AI met with a multi hill map, and I have a bug somewhere. Thanks for the help tho! :) 2011-10-25T20:30:20 why in the tutorial they are doing this now.. TreeSet sortedFood = new TreeSet(ants.getFoodTiles()); 2011-10-25T20:30:28 *** yoden has joined #aichallenge 2011-10-25T20:30:55 Anyone on best way to launch .replay files? 2011-10-25T20:31:20 Tile is appearently sorted now, something about the hashCode() as a comparable.. 2011-10-25T20:31:44 What the first() and last() elements denote now, grom358, I'm afraid I don't know 2011-10-25T20:32:07 yeah but why are they sorting the tiles ?? 2011-10-25T20:32:13 I suppose, maybe you are gettign the closest food? 2011-10-25T20:32:26 As sortedFood.first() 2011-10-25T20:32:38 nah they just loop over the all 2011-10-25T20:32:57 ? 2011-10-25T20:33:03 they get the distance between the food and ants and store the routes and sort the routes 2011-10-25T20:33:12 so the sort on the food and ants seems pointless to me 2011-10-25T20:34:13 Dunno, maybe they are trying to match closest food-ant pairs and give you those routes 2011-10-25T20:34:44 mmmm 2011-10-25T20:35:08 amstan: can you look at the ant paste colors as well? 2011-10-25T20:35:08 *** swamp has quit IRC (Quit: leaving) 2011-10-25T20:35:53 McLeopold: looks good 2011-10-25T20:36:07 *** doricatron has quit IRC (Quit: Page closed) 2011-10-25T20:36:16 *** tncardoso has quit IRC (Quit: bye) 2011-10-25T20:37:37 *** bearoff has joined #aichallenge 2011-10-25T20:38:00 amstan: also, check out JesusBot http://paste.aichallenge.org/jEZT2/ :) 2011-10-25T20:38:03 Why I can no haz replay?! --html "blahahha.html" NO. .replay file yes? "java -jar visualizer.jar *.replay" NO.!! Plz halp 2011-10-25T20:38:08 ah, so ants that die can report their own death that turn 2011-10-25T20:38:27 McLeopold: how did you get that? 2011-10-25T20:39:42 1 more question for y'all: anyone know what the memory limit is? my data structures should use ~300 mb max 2011-10-25T20:39:53 1gb I believe. 2011-10-25T20:40:13 On the official server. Dunno about each TCP server. 2011-10-25T20:40:23 uh, TCP can't see how much resources you're using 2011-10-25T20:40:25 Jesusbot is so great 2011-10-25T20:40:36 @seen mleise 2011-10-25T20:40:36 rmmh: mleise was last seen in #aichallenge 18 hours, 46 minutes, and 4 seconds ago: it always has the best order right in front of the list 2011-10-25T20:40:44 on official it's minimum 1.5GB 2011-10-25T20:41:19 http://aichallenge.org/forums/viewtopic.php?f=24&t=1370&p=9860&hilit=memory#p9860 2011-10-25T20:41:24 McLeopold: any chance of merging my logarithmic graph? 2011-10-25T20:42:02 rmmh, duh my b 2011-10-25T20:45:01 how do these tcp servers work? do they upload your code? or is it just a i/o stream from comp to server? 2011-10-25T20:45:14 *** xar0l has quit IRC (Quit: Page closed) 2011-10-25T20:45:19 *** antonh has quit IRC (Ping timeout: 244 seconds) 2011-10-25T20:45:21 i/o 2011-10-25T20:48:08 *** antonh has joined #aichallenge 2011-10-25T20:49:01 rmmh: it's on the ant paste site, but I'm using sqrt 2011-10-25T20:49:22 McLeopold: ah, how does sqrt behave? 2011-10-25T20:53:00 I like the dirt colored background for the graph 2011-10-25T20:53:49 any official wording anywhere, on how much we're allowed to share/discuss ideas, and share code, etc. 2011-10-25T20:54:22 Hey burny 2011-10-25T20:55:29 *** Relax has quit IRC (Quit: Page closed) 2011-10-25T20:55:58 in particular, one of my friends and me, both plan on competing in this.. and neither of us much cares for the input/output design, but we don't want to re-do it twice either 2011-10-25T20:56:57 *** tdubellz is now known as tdubz 2011-10-25T20:57:08 *** tdubz is now known as tdubellz 2011-10-25T20:58:14 if I want to include a library, does it need to be source code? 2011-10-25T20:58:18 or can i include a jar? 2011-10-25T20:59:00 *** bearoff has left #aichallenge 2011-10-25T20:59:11 *** Six has joined #aichallenge 2011-10-25T20:59:48 *** Puj has quit IRC (Ping timeout: 260 seconds) 2011-10-25T21:00:45 *** spawn has joined #aichallenge 2011-10-25T21:01:11 *** spawn is now known as Guest17743 2011-10-25T21:02:01 *** tan-d42 has quit IRC (Quit: tan-d42) 2011-10-25T21:03:14 burny: I'm sticking to python for a bit; turns out that the main problem was some badly indented sample code xD 2011-10-25T21:03:30 :-) 2011-10-25T21:03:47 *** danielharan has joined #aichallenge 2011-10-25T21:03:50 I'll try to improve the code once I think I have a stable bot on my own 2011-10-25T21:04:00 * avdg is really focussed on his bot 2011-10-25T21:04:24 I got an A* for pathfinding going.. but getting status timeouts :( 2011-10-25T21:04:42 grom358: did you benchmark? 2011-10-25T21:05:45 *** tassadar has quit IRC (Quit: Page closed) 2011-10-25T21:05:50 * avdg has a lot of undefined objects :/ 2011-10-25T21:06:01 *** fr3tles5 has left #aichallenge 2011-10-25T21:06:06 avdg: no.. 2011-10-25T21:06:42 problem is I pathfind every turn in order to get around unseen tiles 2011-10-25T21:07:12 *** pvarga has quit IRC (Ping timeout: 240 seconds) 2011-10-25T21:07:42 *** danielharan has quit IRC (Remote host closed the connection) 2011-10-25T21:08:39 *** he_the_great has joined #aichallenge 2011-10-25T21:09:24 kk.. added in logging on how long the pathfind is taking 2011-10-25T21:09:35 its only saying 0 or 1 millisecond 2011-10-25T21:10:56 *** canadiancow|work has joined #aichallenge 2011-10-25T21:12:16 gggrrr.. seriously.. this makes no sense 2011-10-25T21:13:04 all I'm doing at the moment is A* to visible food and my logging indicates that its fast for couple of turns.. then I just get a timeout 2011-10-25T21:13:56 It's not timing out because you have more ants and more food on the screen, so more things to search? 2011-10-25T21:14:20 and food durther away 2011-10-25T21:14:25 and impossible to get to? 2011-10-25T21:14:41 no I'm testing on the tutorial map. only have couple of ants 2011-10-25T21:15:48 by impossible to get to, I mean.. that you can't see a path to it, but can see the food? 2011-10-25T21:16:02 Does each turn take progressively longer until eventually you overuse your time? Or is it low, low, low, low, timeout? 2011-10-25T21:16:04 crashing on turn 12.. and my benchmark says I only took 5ms 2011-10-25T21:16:16 its low, low, low, low, timeout 2011-10-25T21:16:59 eg.. INFO: #12,FOOD TIME: 5 2011-10-25T21:17:08 *** TheDigitalNinja has joined #aichallenge 2011-10-25T21:17:19 meaning I took 5miliseconds to path and order moves to food on turn 12 2011-10-25T21:18:20 oh, so you completed pathfinding and still timed out? 2011-10-25T21:18:36 *** pvarga has joined #aichallenge 2011-10-25T21:18:47 yeah.. it looks like it timesout somewhere on turn #13 2011-10-25T21:19:11 don't get why 2011-10-25T21:19:22 debug through turn 13? 2011-10-25T21:19:46 yeah.. gonna add more logging.. lets see 2011-10-25T21:20:20 *** bobbydroptable has joined #aichallenge 2011-10-25T21:20:26 wow major unheaval in ranks 2011-10-25T21:21:49 *** canadiancow has joined #aichallenge 2011-10-25T21:21:49 *** canadiancow|work has quit IRC (Disconnected by services) 2011-10-25T21:22:13 *** Guest17743 has quit IRC (Ping timeout: 265 seconds) 2011-10-25T21:22:22 *** canadiancow|work has joined #aichallenge 2011-10-25T21:26:08 *** canadiancow has quit IRC (Ping timeout: 244 seconds) 2011-10-25T21:28:51 what exactly does the 'skill' mean, in the rankings? 2011-10-25T21:29:15 burny: the trueskill score thing 2011-10-25T21:29:16 burny: I believe it's your TrueSkill score. 2011-10-25T21:29:26 which is? 2011-10-25T21:29:34 burny: A number between 0 and 100. 2011-10-25T21:30:17 To be less snarky, it's an estimate of how "good" a player is using the TrueSkill algorithm. 2011-10-25T21:30:21 http://research.microsoft.com/en-us/projects/trueskill/faq.aspx 2011-10-25T21:30:32 It doesn't mean "percentage of games you've won" or anything so intuitive 2011-10-25T21:30:34 eww, microsoft 2011-10-25T21:30:39 *withdraws from the contest* 2011-10-25T21:30:59 cyphase: how do you think i feel? 2011-10-25T21:31:16 lol 2011-10-25T21:31:43 It turns out Microsoft has some experience ranking skill of players. 2011-10-25T21:32:01 whodu thunk 2011-10-25T21:38:19 erm.. so it's basically similar to ELO, except, it's auto normalized to a 1-100 range, so it has no meangful value, or 'unit'? 2011-10-25T21:38:38 *** edcba has quit IRC (Read error: Operation timed out) 2011-10-25T21:38:39 *** Craklyn has quit IRC (Ping timeout: 265 seconds) 2011-10-25T21:39:03 It is not ELO, but it serves a similar purpose. 2011-10-25T21:39:18 *** edcba has joined #aichallenge 2011-10-25T21:39:48 in ELO i think you lose points for games lost, and gain points for games won 2011-10-25T21:40:03 in trueskill, you don't lose points if you lose to another player who was obviously better than you 2011-10-25T21:40:08 I don't know the details of Elo, but with TrueSkill, Each player has a pair (mu, sigma) which represent's the player's average skill (mu) and the deviation in that (sigma). 2011-10-25T21:40:21 but you lose points if you lost to someone who you should have beaten 2011-10-25T21:40:24 *** g3 has joined #aichallenge 2011-10-25T21:40:43 You can read the documentaiton on TrueSkill, but basically as its expectations about how well you will do are confirmed, mu goes to your "actual skill" and sigma goes to zero. 2011-10-25T21:40:48 *** g3 is now known as skdjfsdkjfn 2011-10-25T21:40:50 s/documentaiton/documentation/ 2011-10-25T21:41:13 ya, but the 'number', is meaningless, yes? 2011-10-25T21:41:31 where-as elo, the number has an actual value/unit 2011-10-25T21:41:50 *** skdjfsdkjfn has left #aichallenge 2011-10-25T21:42:58 i'm guessing its more of a probability? 2011-10-25T21:44:33 *** tassadar has joined #aichallenge 2011-10-25T21:44:41 burny: No, it doesn't have a meaningful unit as far as I know. 2011-10-25T21:44:51 But I do know that TrueSkill is rather effective in practice. 2011-10-25T21:45:02 I've heard Elo doesn't work well with >2 players per game, but haven't read too much into it. 2011-10-25T21:45:20 in the game settings, it says foodrate = 2,8. which one should we use for testing? 2011-10-25T21:46:13 *** rxc has joined #aichallenge 2011-10-25T21:46:27 tassadar: test them both ;-) 2011-10-25T21:46:53 *** danielharan has joined #aichallenge 2011-10-25T21:47:08 at least, try to provide as much scenarios you can create and test your bot if it can handle these situations well 2011-10-25T21:47:37 *** bmh has joined #aichallenge 2011-10-25T21:47:38 Out of curiosity, what are the units on Elo score? 2011-10-25T21:47:42 *** bmh has quit IRC (Changing host) 2011-10-25T21:47:42 *** bmh has joined #aichallenge 2011-10-25T21:47:50 jbroman: it's unitless 2011-10-25T21:48:00 I forget exactly, but something along the lines of.. 2011-10-25T21:48:19 any contest admins around? I want to know if a particular strategy is considered unfair. 2011-10-25T21:48:22 each 100score difference, is a 2:1 expectation of winning 2011-10-25T21:48:32 bmh: yes? 2011-10-25T21:49:00 *** GGG has joined #aichallenge 2011-10-25T21:49:19 pretty much any strategy is fair, aside from communicating with other bots 2011-10-25T21:49:21 *** ig has quit IRC (Read error: Connection reset by peer) 2011-10-25T21:49:27 can anyone suggest to me a strategy for deciding the best location to scout at the beginning other than "wander around randomly" 2011-10-25T21:49:42 bobbydroptable: solve the canadian traveller problem ;) 2011-10-25T21:49:47 i've been thinking about it for 2 days and i've thought about waypointing (but i have no idea how to decide the best place to waypoint) 2011-10-25T21:50:10 and i've looked into navigation meshes (but i can't decide how to set the regions especially with an unknown map) 2011-10-25T21:50:19 bobby.. use probability, and determine which path, is going to generate the best proability of finding food 2011-10-25T21:50:48 taking into account how likely you are to get food, how long it will take to get the food, how likely you are to die before you get it, etc. 2011-10-25T21:50:57 canadian traveller: that's exactly the problem ==" looking into it 2011-10-25T21:51:23 burny that's alot of things to take into account ==" 2011-10-25T21:51:41 yup :) 2011-10-25T21:53:28 :( 2011-10-25T21:53:56 I have a concern about the scoring system, any admins who are familiar with how exactly it works, care to discuss in /msg? 2011-10-25T21:54:40 burny: sure? 2011-10-25T21:55:27 how big normally is the delay between your turn end to your next turn start? 2011-10-25T21:55:37 bobbydroptable: This problem actually appears to be a hellish variant of the CTP -- multi-agent CTP 2011-10-25T21:56:00 it doesn't have to be though 2011-10-25T21:56:03 i think 2011-10-25T21:56:30 what i've been trying to do is to build a graph across the entire map so that i don't need to tell my ants where to go, they decide based on weights on the nodes 2011-10-25T21:56:48 but I can't do even distribution of nodes because there will be areas where the nodes fall on water blocks 2011-10-25T21:56:57 and i can't figure out how to build the graph dynamically as I traverse the environment 2011-10-25T21:57:43 so its not a multi-agent CTP in that there is only 1 graph I am building. this abstracts away the direction decision making from the ant agents 2011-10-25T21:57:57 and from the sounds of it, this hasn't been solved yet? 2011-10-25T21:58:12 at least, still NP problem? 2011-10-25T21:59:05 you could create repulsion and attraction points in the map, that modify the A* weight it puts on a movement depending on your distance to these points 2011-10-25T21:59:18 so that ants naturally will take paths that go through the attraction points 2011-10-25T21:59:23 and avoid the repulsion points 2011-10-25T21:59:56 Draivin that's a fine idea 2011-10-25T22:00:02 but where do i put the points? 2011-10-25T22:00:15 how do i decide the best place to put the point 2011-10-25T22:00:48 they could be dinamic, put attraction points in enemy hills, and adjust the repulsion points based on the number of enemy ants in an area 2011-10-25T22:00:49 that's essentially the problem i am having :) 2011-10-25T22:01:07 but at the beginning none of these things are in play 2011-10-25T22:01:15 there may be a few food around 2011-10-25T22:01:28 but if there's no food around there would be no goal 2011-10-25T22:01:44 you could assign a task to a ant 2011-10-25T22:01:50 like when there is a unclaimed food 2011-10-25T22:01:55 you find the nearest and 2011-10-25T22:01:55 bah, making food collection maximally efficient is hard 2011-10-25T22:01:58 ant assign it to get thtat food 2011-10-25T22:02:05 that* 2011-10-25T22:02:08 ant* 2011-10-25T22:02:12 and* 2011-10-25T22:02:20 *** tassadar has quit IRC (Quit: Page closed) 2011-10-25T22:02:20 and when there is no food, what do the ants do? 2011-10-25T22:02:27 other than "wander aimlessly" 2011-10-25T22:02:29 Whatever you tell them to do, of course! 2011-10-25T22:02:39 what do i tell them to do!? 2011-10-25T22:02:40 lol 2011-10-25T22:02:44 bobbydroptable: you could try killing enemy ants. 2011-10-25T22:02:49 *** Islacrusez has quit IRC (Ping timeout: 265 seconds) 2011-10-25T22:02:51 if it can listen of course ;-) 2011-10-25T22:03:03 okay okay lets go back to the beginning 2011-10-25T22:03:09 you start off on a hill 2011-10-25T22:03:14 you only see water blocks around you 2011-10-25T22:03:21 ruuun 2011-10-25T22:03:32 there's no food 2011-10-25T22:03:40 panic? 2011-10-25T22:03:40 set a pathfinding to a unknown point in the map 2011-10-25T22:03:46 and change it dinamically 2011-10-25T22:03:48 as you discover 2011-10-25T22:03:49 the map 2011-10-25T22:04:07 * avdg has an other secret plan 2011-10-25T22:04:16 okay so imagine there's water blocks surrounding N,S,W 2011-10-25T22:04:21 so obviously to me, i need to go E 2011-10-25T22:04:26 not really 2011-10-25T22:04:36 no? 2011-10-25T22:04:38 if you set a path to W 2011-10-25T22:04:42 with a good pathfinding 2011-10-25T22:04:49 it will try to wander around the water 2011-10-25T22:04:52 and explore automatically 2011-10-25T22:05:04 sounds expensive 2011-10-25T22:05:18 thats just while you are exploring blocks 2011-10-25T22:05:21 * avdg doesn't like pathfinders on unknown areas 2011-10-25T22:05:22 and creating your map vision 2011-10-25T22:05:28 avdg: me too == 2011-10-25T22:05:31 besides 2011-10-25T22:05:36 you could limit the pathfinding 2011-10-25T22:05:42 so it only gets the next 5 blocks or so 2011-10-25T22:05:46 which could give you suboptimal results. 2011-10-25T22:05:48 per run 2011-10-25T22:05:48 *** ig has joined #aichallenge 2011-10-25T22:06:00 and what I'm planning to do 2011-10-25T22:06:08 well the problem is 2011-10-25T22:06:12 mine goes probably like 10 tiles 2011-10-25T22:06:14 if you do a limit on A* 2011-10-25T22:06:24 you won't know if the last point you got to was the best subpath 2011-10-25T22:06:49 it will be the best subpath you could get 2011-10-25T22:06:49 if you limit it to your view 2011-10-25T22:06:50 at least, that's what i found when i tried it anyway 2011-10-25T22:06:52 in those circunstances 2011-10-25T22:06:56 get* 2011-10-25T22:07:39 it was quite funny when i tried it 2011-10-25T22:07:57 on one block, it found the "best" path to be forward, and then on the next turn, it found the best path to be behind it. 2011-10-25T22:08:24 *sigh* i will mull over it more 2011-10-25T22:08:26 more pen paper; 2011-10-25T22:08:28 haha 2011-10-25T22:08:34 more visualization ;-) 2011-10-25T22:08:40 oh i've tried lol 2011-10-25T22:08:50 i've tried to see if i could do a flooding to detect regions 2011-10-25T22:09:04 *** SuperMole has joined #aichallenge 2011-10-25T22:09:12 in the lack of options, just stick to a wall and keep following it 2011-10-25T22:09:15 you will explore eventually 2011-10-25T22:09:16 =) 2011-10-25T22:09:25 :p 2011-10-25T22:09:31 only in maze maps :) 2011-10-25T22:09:34 lol 2011-10-25T22:09:38 *** jesionaj has quit IRC (Ping timeout: 276 seconds) 2011-10-25T22:09:54 *** SuperMole has left #aichallenge 2011-10-25T22:10:19 what do you do on land? :-) 2011-10-25T22:10:29 wander randomly 2011-10-25T22:10:29 lol 2011-10-25T22:10:30 haha 2011-10-25T22:10:32 or what if the only piece of water is a rectangle? 2011-10-25T22:10:44 exactly! ==" 2011-10-25T22:11:00 *** SuperMole has joined #aichallenge 2011-10-25T22:11:26 *** SuperMole has left #aichallenge 2011-10-25T22:12:21 *** djr_ has joined #aichallenge 2011-10-25T22:12:32 *** swamp has joined #aichallenge 2011-10-25T22:12:34 and I don't think I will put a lot of probability in my ai 2011-10-25T22:12:42 just food for thought, but you could also implement a multi-tier A*, at the most basic level it treats larger squares of land as one tile 2011-10-25T22:12:45 and the closer you get 2011-10-25T22:12:52 the more detailed the pathfinding gets 2011-10-25T22:12:52 draivin that is what i want to do ==" 2011-10-25T22:12:55 *** Craklyn has joined #aichallenge 2011-10-25T22:13:12 but larger squares lead to problems. 2011-10-25T22:13:14 bobbydroptable: I've solved with stress balls :p 2011-10-25T22:13:21 *it with 2011-10-25T22:13:26 literally or figuratively 2011-10-25T22:13:27 lol 2011-10-25T22:13:31 both 2011-10-25T22:13:56 hmmm, stress balls and maps, what could that be :-) 2011-10-25T22:14:15 don't understand :( 2011-10-25T22:14:28 i'm certainly stressed 2011-10-25T22:14:47 (hint: you can nip them and force shapes) 2011-10-25T22:14:48 more thoughts, you could implement a 'squad' as you get more ants, and treat each squad as a single entity, the ants will stick to the squad, and you could do a single pathfinding for a whole squad 2011-10-25T22:14:58 *** swamp has left #aichallenge 2011-10-25T22:15:02 draivin: we're still thinking about late game 2011-10-25T22:15:07 my problem is early game related 2011-10-25T22:15:42 everything in the late game can be solved, but only if i have a path to begin with. 2011-10-25T22:15:56 and at the beginning there is no path 2011-10-25T22:16:10 even the tutorials have a good hint how to start 2011-10-25T22:16:18 i didn't like it 2011-10-25T22:17:30 *** swamp has joined #aichallenge 2011-10-25T22:18:06 meh, my latest log contains 21451 times the word undefined 2011-10-25T22:18:09 I don't like it 2011-10-25T22:18:20 lol 2011-10-25T22:18:22 lol² 2011-10-25T22:18:39 avdg i think we're both on the same wavelength. we both want more control over what our ants do, and less randomness 2011-10-25T22:19:00 meh, I hate having controls 2011-10-25T22:19:15 lol ok 2011-10-25T22:19:25 it sound more like a system where individual ants gets more rights 2011-10-25T22:19:32 thats no-no for teamwork 2011-10-25T22:19:39 and cpu 2011-10-25T22:19:40 don't you have 100% complete control? 2011-10-25T22:20:18 nope, the algorithm should have complete control about the situation 2011-10-25T22:20:31 I just dictate and teach the algo :p 2011-10-25T22:20:52 genetic algo? :D 2011-10-25T22:21:00 meh 2011-10-25T22:21:25 to be honest, I have some inspiration from the google search algo 2011-10-25T22:21:34 though its not really the same 2011-10-25T22:23:13 avdg: doing something with markov chains? 2011-10-25T22:23:20 nah 2011-10-25T22:23:25 I don't like technical terms 2011-10-25T22:23:49 *** jesionaj has joined #aichallenge 2011-10-25T22:23:49 bmh: tell me what you'll get, i'm curious 2011-10-25T22:23:51 it's too mainstream 2011-10-25T22:23:51 lol 2011-10-25T22:23:58 it only shows half-baked algorithms which doesn't get sense many times 2011-10-25T22:24:09 I have only 1 problem, I'm very slow in coding :/ 2011-10-25T22:24:12 amstan: tell you what I'll get where? 2011-10-25T22:24:19 i have 1 problem, i can't solve problems. 2011-10-25T22:24:22 bmh: with your strategy 2011-10-25T22:24:25 XD 2011-10-25T22:24:34 amstan: I'll let you know when I implement it 2011-10-25T22:24:36 just solve that problem, and you'll be set :P 2011-10-25T22:24:38 bobbydroptable: are you making a recursive solution? 2011-10-25T22:24:43 I have a problem like avdg and bobbydroptable: I'm a fool. 2011-10-25T22:24:44 nope 2011-10-25T22:24:47 all iterative. 2011-10-25T22:25:01 too much overhead with recursion 2011-10-25T22:25:10 nah, I mean your thinking 2011-10-25T22:25:11 (iteration is recursion.) 2011-10-25T22:25:30 depends on your stack 2011-10-25T22:25:40 well, a function call can be a recursion as well 2011-10-25T22:25:45 my thinking is... build different components of what i need. 2011-10-25T22:25:55 i've got a ant tracker now 2011-10-25T22:26:01 and i've got a graph 2011-10-25T22:26:04 amstan: I'm going to start nerdraging soon, let's talk about ants :) 2011-10-25T22:26:09 I'm fighting hard for a stable core with some extra features 2011-10-25T22:26:13 bmh: lol sorry 2011-10-25T22:26:25 bobbydroptable: yeah… ant tracking is a pain. Too bad ants don't just have unique ids. 2011-10-25T22:26:37 bmh: I am solving that problem atm 2011-10-25T22:26:38 Update my own state!? What!? 2011-10-25T22:26:41 but why would you need to track them? 2011-10-25T22:26:49 assign tasks/goals 2011-10-25T22:26:56 or other things 2011-10-25T22:27:04 its just easier to program 2011-10-25T22:27:06 :-) 2011-10-25T22:27:11 like which direction it is biased towards when scouting. 2011-10-25T22:28:17 i think what i will do is to do a BFS in all 4 directions, see which one gets me the furthest. 2011-10-25T22:28:24 then place a node there. 2011-10-25T22:28:26 is 4 enough? 2011-10-25T22:28:45 no need for diagonals i think... 2011-10-25T22:28:49 Hey, is anybody using potential field for his bot? 2011-10-25T22:29:02 i don't even know what that means :( 2011-10-25T22:30:08 *** Knekkebjoern has joined #aichallenge 2011-10-25T22:30:57 a1k0n: thinking about that "tracking down enemies" idea, i think it might just be alright to keep information at each land location out of sight for the soonest an enemy you previously saw could have reached that location 2011-10-25T22:31:39 the only difference really from doing it probabilistically i think is that it'd weight locations with more access more? 2011-10-25T22:31:56 That's a method to assign a score to each square and your ants move by choosing the highest square/score next to them 2011-10-25T22:32:20 Are players going to hide their good stuff until the very end? Or would you think most players are uploading their best bot so far 2011-10-25T22:32:31 bmh: btw, antimatroid was the designer of the map stuff 2011-10-25T22:32:32 Larose: my first bot did that 2011-10-25T22:32:39 but ran into problems 2011-10-25T22:32:45 wasn't sure if it was the best idea ==" 2011-10-25T22:32:45 i'm calling you good guy greg now and yes most people just upload their best bot 2011-10-25T22:32:50 although some may just test on tcp 2011-10-25T22:32:52 Like what problems? 2011-10-25T22:32:52 meh, I revealed a lot of my ideas, but mostly not on a clear way 2011-10-25T22:32:57 amstan: Thanks 2011-10-25T22:33:07 well 2011-10-25T22:33:13 in multi hive maps 2011-10-25T22:33:23 GGG: I'm going to post my bot once I 1) recruit my team, 2) write it 2011-10-25T22:33:38 well what i did was added to the score where my ants were so that they wouldn't go back 2011-10-25T22:33:52 but when it met another ant, it got confused with the other ants "path" and turned back 2011-10-25T22:33:56 lol 2011-10-25T22:34:02 bmh: don't actually share your bot code till after the contest please :) 2011-10-25T22:34:09 you probably didn't mean that, just saying 2011-10-25T22:34:17 Larose: I am doing this right now for my "explorer" ants 2011-10-25T22:34:19 Does anyone know the address of a tcp test server? 2011-10-25T22:34:19 antimatroid: %s/post/submit 2011-10-25T22:34:19 antimatroid: that's the GGG caption 2011-10-25T22:34:41 bmh: what were you saying about maps? 2011-10-25T22:34:50 Nescio: http://ants.fluxid.pl/ ? 2011-10-25T22:34:51 antimatroid: nothing. secrets :) 2011-10-25T22:35:30 meh, lets play poker :-) 2011-10-25T22:35:40 @tcp 2011-10-25T22:35:41 amstan: I do not know about 'tcp', but I do know about these similar topics: 'topic' 2011-10-25T22:35:49 @learn tcp as http://ants.fluxid.pl/ 2011-10-25T22:35:49 amstan: OK 2011-10-25T22:36:03 Larose: Have you tried potential fields as well? 2011-10-25T22:37:57 Craklyn: Yes, but I don't know how to navigate in maze map 2011-10-25T22:38:10 amstan: is anything pointing at the planet wars website? 2011-10-25T22:38:19 nope 2011-10-25T22:38:23 Larose: A*/Dijkstra? 2011-10-25T22:38:25 *** Exception has joined #aichallenge 2011-10-25T22:38:53 antimatroid: if you just want to read something off it, you can get the ip of jpcameron.com, then locally point ai-contest.com to it 2011-10-25T22:38:57 by editing your hosts file 2011-10-25T22:39:00 Is there a standard way to do a "potential field"? I made an ad-hoc solution which didn't really work and I'm some of the parameters I created trying to see if I can get it to function... 2011-10-25T22:39:09 I'll describe my method if you're interested :P 2011-10-25T22:39:11 amstan: someone just sent out an email about it 2011-10-25T22:39:20 plus i'd been curious a couple of times, but i don't mind that much atm 2011-10-25T22:39:23 Craklyn: I'm interested :) 2011-10-25T22:39:53 I made a 2d array of integers which corresponds to the map. Each integer begins at -1. 2011-10-25T22:40:12 Each turn, the map space which corresponds to the square the ants are standing on, I set the value to the current turn 2011-10-25T22:40:14 Should I expect the time per turn to vary? 2011-10-25T22:40:45 For all the squares the ants can see, I set the value to MAX(tile's current value, turnNumber - distance2(ant, tile)) 2011-10-25T22:41:02 bmh: Well, because food attracks my ant, they always hit a wall.. 2011-10-25T22:41:19 (Actualy, this is centered on the square the ant came from) 2011-10-25T22:41:21 attracts* 2011-10-25T22:41:23 Larose: Don't include water in your search graph. 2011-10-25T22:41:36 The ant then simply walks downhill.. 2011-10-25T22:42:21 Craklyn: That's cute. Can you link to an example of it in action? 2011-10-25T22:42:48 I literally made this iteration <5 minutes ago. Am double checking to see if there's bugs, etc 2011-10-25T22:43:12 *** jesionaj has quit IRC (Ping timeout: 240 seconds) 2011-10-25T22:43:13 (The last iterations tended to get stuck going in circles) 2011-10-25T22:44:10 Craklyn: I don't really get the MAX( ... ) 2011-10-25T22:44:11 I'm using A* search for all of my ants, but it starts timing out when I get over 20 ants 2011-10-25T22:44:22 How can I optimize the pathfinding algorithm? 2011-10-25T22:44:43 Exception: Try hierarchical A* 2011-10-25T22:45:17 If you think about the way this algorithm works, I am basically building a hill around all the ants 2011-10-25T22:45:18 For the moment, water repulses and food attracts (according to the manhattan distance) 2011-10-25T22:45:32 And then requiring the ants fall down hill continuously 2011-10-25T22:45:52 MAX ensures that I don't dig a hole for the ant to trap itself into.. 2011-10-25T22:46:34 I'm not sure if the figurative language is helpful or not :P 2011-10-25T22:46:46 I was considering using a heavily weighted A* algorithm, I think the distances are short enough that it doesn't matter if isn't a completely optimal shortest distance. 2011-10-25T22:46:48 I think I get it 2011-10-25T22:47:00 It's to prevent to get stuck in a local minima? 2011-10-25T22:47:02 this is driving me nuts... doing the pathfinding is so hard 2011-10-25T22:47:40 my nice A* can't handle the big maps.. status timeouts :( 2011-10-25T22:47:42 It's to prevent the code from making the field a minimum at a place you recently measured well 2011-10-25T22:47:52 doing it all together at once is hard ;-) 2011-10-25T22:48:31 I increase the map's potential for places I know well. The center of an ant gets a value of (current turn). As you get further from an ant, the value goes up to (current turn - distance2) 2011-10-25T22:49:09 However, if you measured it very very well two turns ago, you "want" the value to be (current turn - 2). If you didn't have the max, it'd be downgraded to (current turn - 4). 2011-10-25T22:49:44 grom358: are you pathfinding at every turn? 2011-10-25T22:49:56 It's an ad-hoc solution. I'm not saying it's the proper way to weight it :P I was hoping to discuss it because maybe I'm doing something dumb.. 2011-10-25T22:51:35 Craklyn: Maybe I didn't read the same papers as you did, but I can't figure out how you compute the attractive/repulsion of the water/food/ants/hill 2011-10-25T22:52:00 This is a purely "explorer" ant. He's only interested in revealing new territory 2011-10-25T22:52:09 Oh, ok 2011-10-25T22:52:10 I'm making different discrete ant goals right now 2011-10-25T22:52:28 Later, I can just sum potentials to get a superpotential :) 2011-10-25T22:52:56 grom358: if you're timing out I would suggest doing this: Keep a queue of ants and their marching orders, sorted by age 2011-10-25T22:53:07 I didn't read any papers on this subject. Is this the general way a person does a "potential field" in AI? 2011-10-25T22:53:11 grom358: Then do path-finding for the unordered ants until you run out of time 2011-10-25T22:53:42 bmh: yeah I do it everytime.. yeah I'll have todo a queue 2011-10-25T22:53:55 That's a good idea, maybe I can use potential field only for my ants explorer 2011-10-25T22:54:40 * avdg read ants as "ants explorer" as browser for ants 2011-10-25T22:54:55 *** swamp has left #aichallenge 2011-10-25T22:56:20 lol 2011-10-25T22:58:20 Right now, my ants get often stuck in local minima (using manhattan distance for food/water). So i though that I could do a BFS from the food with a depth limit for the attraction. But this doesn't seem efficient 2011-10-25T23:00:58 Larose: That's what I currently do for my "eater" ants. I do a single BFS from all foods on the map up to the max distance any ant can see. I then have all my eater ants walk downhill toward their nearest food. It's super quick to calculate, but the downside is that I sometimes get multiple ants going toward the same food. 2011-10-25T23:01:05 *** swamp has joined #aichallenge 2011-10-25T23:01:40 *** swamp has left #aichallenge 2011-10-25T23:03:32 craklyn why don't you just sort ants by distance then tell them to gather 2011-10-25T23:03:33 *** svujic has joined #aichallenge 2011-10-25T23:03:47 *** svujic has left #aichallenge 2011-10-25T23:03:58 For example, A _ _ F _ _ A 2011-10-25T23:04:21 Both ants are equidistance from the food, so they both go downhill toward the food 2011-10-25T23:04:22 A _ _ F _ _ A _ _ F 2011-10-25T23:04:43 Both ants are equidistant from a food. I'd have to do more calculations for them to find out they can each chase down their own food 2011-10-25T23:04:59 ah yeh 2011-10-25T23:05:05 I tried to calculate a super fast way to gather food so that I'd potentially have time to take care of more difficult calculations, such as combat 2011-10-25T23:06:12 How do you guys handle battling? 2011-10-25T23:06:14 *** avdg has quit IRC (Quit: Leaving.) 2011-10-25T23:07:13 And how do you prioritize gathering food with battling? 2011-10-25T23:07:22 I haven't begun battle yet :( 2011-10-25T23:07:36 Do you know about Min-Max games? 2011-10-25T23:07:41 ti have a plan regarding that :D 2011-10-25T23:07:42 *** avdg has joined #aichallenge 2011-10-25T23:08:36 :( 2011-10-25T23:08:43 *** svujic has joined #aichallenge 2011-10-25T23:09:11 For combat, I was thinking I'd run a MIN-MAX algorithm with only one round of look-ahead, or two if time permits 2011-10-25T23:09:25 For pathfinding, I calculated the shortest path between every walkable square. 2011-10-25T23:09:26 But with the huge number of possible moves, I assume I can only look one round ahead 2011-10-25T23:09:43 but I die to other ants 2011-10-25T23:10:13 because usually other ants are in the way of the path 2011-10-25T23:11:23 As a super simple step, you could have the ants stop going if they're going to walk into a square where they could die.. 2011-10-25T23:12:52 Hmm 2011-10-25T23:13:10 The batting part is annoying me 2011-10-25T23:13:16 battling* 2011-10-25T23:13:29 Yeah, me too. I can't intuit how combat works in this game yet 2011-10-25T23:13:51 Does anyone have a strategy for battling? 2011-10-25T23:14:00 did you read the long description where you should count enemies for each ant and look for lowered number enemies? 2011-10-25T23:14:02 I do 2011-10-25T23:14:11 Exception: don't die 2011-10-25T23:14:29 yeah I know 2011-10-25T23:14:36 but is there a different way to play the game? 2011-10-25T23:14:50 ya.. 2011-10-25T23:15:00 balance killing, vs not-dieing, vs controlling teritory 2011-10-25T23:15:02 *** svujic has left #aichallenge 2011-10-25T23:15:03 like don't collect food and try to kill all the enemy ants by splitting them up 2011-10-25T23:15:12 *** svujic has joined #aichallenge 2011-10-25T23:15:13 #1 priority is capturing hills 2011-10-25T23:15:22 craklyn: even if you partition ants into sets of those that can affect each other i think you'll be struggling to run minimax 2011-10-25T23:15:23 burny: how would you balance those 3? 2011-10-25T23:15:53 Antimatroid: Even for only one turn in the future? 2011-10-25T23:15:56 in clever complicated ways 2011-10-25T23:16:13 each ant has up to 5 moves, that's up to 5^n states for n ants in your tree 2011-10-25T23:16:26 terminal states even 2011-10-25T23:16:30 could also look at food-history in the area 2011-10-25T23:16:35 meh, I found a bot that has a very same behavior as my current submitted bot :p 2011-10-25T23:16:46 *** Bleda has quit IRC (Ping timeout: 265 seconds) 2011-10-25T23:16:47 and how close it is to your hill, or a potential enemy hill 2011-10-25T23:16:58 hmm 2011-10-25T23:17:06 *** svujic has left #aichallenge 2011-10-25T23:17:33 I think you have to balance the number of food gatherers with the number of hills you control 2011-10-25T23:17:41 for battle i'm considering generating information around ants that can battle as to how many enemys could be in range of each possible location to move into at the end of the turn and maybe use a decision tree or something 2011-10-25T23:17:50 *** svujic has joined #aichallenge 2011-10-25T23:17:53 how many enemies in the worst case 2011-10-25T23:18:09 yes, but you only have limited vision 2011-10-25T23:18:44 and what if you're attacking into a chokepoint? 2011-10-25T23:18:55 like 2011-10-25T23:19:01 %%%%%%% 2011-10-25T23:19:01 i was thinking you could precompute minimax in a file, and load into memory at beginning lol 2011-10-25T23:19:08 lol 2011-10-25T23:19:19 bobbydroptable: lol except the maps and ants moves change 2011-10-25T23:19:19 you can't write to files 2011-10-25T23:19:25 Antimatroid: I agree with your point - if combat range is increased or there's lots of interfaces between your ants and the enemies, then combat is impossible to look even one turn in the future with min-max unless you have a very, very simple way to calculate the "value" of a state 2011-10-25T23:19:28 you aren't going to minimax an entire game, that's more absurd than chess 2011-10-25T23:19:32 no writing. just read 2011-10-25T23:19:36 compute locally and upload 2011-10-25T23:19:36 lol 2011-10-25T23:19:42 lol 2011-10-25T23:19:43 You don't even have complete information 2011-10-25T23:19:51 hmm 2011-10-25T23:20:18 than for* chess 2011-10-25T23:20:30 eg. if enemy has 12 ants there, and I have 5, and I don't think enemy base is near by, and enemy isn't attackign or fleeing, I might want to keep my 5 ants there as decoy, rather than sending them on other missions, or gathering extra troops to attack 2011-10-25T23:20:37 no just for particular situations not the entire board 2011-10-25T23:20:58 I got the pathfinding down 2011-10-25T23:21:04 burny: even with numerical superiority, a good formation can still defeat a bad one 2011-10-25T23:21:14 That's a good point. For small engagements of <5 vs. <5, you can precompute solutions if you're crazy enough.. 2011-10-25T23:21:19 Yes, formation is important 2011-10-25T23:21:36 *** Larose has quit IRC (Ping timeout: 240 seconds) 2011-10-25T23:21:38 But how do you find the optimal formation? 2011-10-25T23:21:41 hmmm that's not a bad idea 2011-10-25T23:21:44 game tree 2011-10-25T23:21:54 There's no "scouting" in this game 2011-10-25T23:22:00 There isn't? 2011-10-25T23:22:12 If you see your opponent ants, then your ant dies 2011-10-25T23:22:19 are we guaranteed that view_distance > attack_distance? 2011-10-25T23:22:22 Depends on relative range 2011-10-25T23:22:27 rmmh: yes 2011-10-25T23:22:36 oh 2011-10-25T23:22:51 if view_dist > attack_dist then it works 2011-10-25T23:23:25 but it doesn't say in the specification if thats always true 2011-10-25T23:23:28 If that weren't the case, then you'd just need to figure out the optimal formation and march your blind units to their deaths... :( Poor guys... 2011-10-25T23:23:44 lol 2011-10-25T23:23:49 aichallenge: McLeopold epsilon * rf3538e6 / (sql/2_generate_matchup.sql sql/opponent.sql): fix empty database matchup creation issue - http://git.io/vyOQmQ 2011-10-25T23:23:49 well 2011-10-25T23:23:51 sounds un-fun 2011-10-25T23:23:55 i believe the optimal formation to be a line. 2011-10-25T23:23:58 no 2011-10-25T23:24:09 optimal formation is an arc 2011-10-25T23:24:33 Lines are vulnerable at corners. Arcs are vulnerable at the tip.. 2011-10-25T23:24:44 do you mean concave or convex 2011-10-25T23:24:50 concave 2011-10-25T23:25:09 Any place your units come to a point, the unit which sticks out is vulnerable.. 2011-10-25T23:25:24 a circle then. 2011-10-25T23:25:25 (Except when you use landscape to block your tips) 2011-10-25T23:25:26 yes 2011-10-25T23:25:26 lol 2011-10-25T23:25:36 thats why you use the terrain 2011-10-25T23:25:42 yeh defensive terrain always helps out 2011-10-25T23:25:57 put the endpoints against the terrain 2011-10-25T23:26:06 with the irregular terrain, it should be possible to make defensive formations that can't be attacked 2011-10-25T23:26:20 yes 2011-10-25T23:26:32 there are a lot of chokepoints on the map 2011-10-25T23:26:39 hmm 2011-10-25T23:26:52 but defense is only part of the puzzle, you only gain points when you kill other hills 2011-10-25T23:27:10 but your enemy can't kill your hills 2011-10-25T23:27:23 yes but your enemy can take out other hills 2011-10-25T23:27:29 while you're stuck there 2011-10-25T23:27:35 yea 2011-10-25T23:27:42 This game is too hard. Let's all agree to code an AI for Go instead? 2011-10-25T23:27:46 lol 2011-10-25T23:27:47 ==" 2011-10-25T23:27:48 lol 2011-10-25T23:27:54 this game is interesting 2011-10-25T23:28:02 ok how about this 2011-10-25T23:28:11 I would love to do ai for go, if I knew atleast 1 other person who was highly interested in it 2011-10-25T23:28:27 you can use squads of ants to block enemy chokes 2011-10-25T23:28:27 i like go 2011-10-25T23:28:28 suck shit at ai though 2011-10-25T23:28:39 exception: that seems to be the idea. 2011-10-25T23:28:40 so they can't get out of their territory 2011-10-25T23:28:48 territory control is key 2011-10-25T23:28:53 yes 2011-10-25T23:29:05 your ai can then determine if it has enough ants to bust through 2011-10-25T23:29:11 um 2011-10-25T23:29:16 or whether it is more prudent to hold gorund 2011-10-25T23:29:23 some places are impossible to bust through 2011-10-25T23:29:38 *** Relax has joined #aichallenge 2011-10-25T23:29:57 actually, all chokes are impossible to bust through given the battle mechanics 2011-10-25T23:30:18 I think it depends on the choke 2011-10-25T23:30:37 Also, it can depend on the direction you're trying to bust through 2011-10-25T23:31:05 *** yoden has quit IRC (Quit: Leaving.) 2011-10-25T23:31:12 If you're in a two-wide tunnel trying to move into a 4-wide tunnel, you're stuck if there are 4 guys standing side by side 2011-10-25T23:31:23 are there any two-wide tunnels? 2011-10-25T23:31:25 If you're in a 4-wide tunnel trying to move into a 2-wide tunnel, you can put two of your units in and trade for two enemy units 2011-10-25T23:31:34 Like, imagine the terrain: 2011-10-25T23:31:41 X o o o o X 2011-10-25T23:31:45 X X o o X X 2011-10-25T23:31:53 X being water, o being open space 2011-10-25T23:32:06 *** pipeep has joined #aichallenge 2011-10-25T23:32:08 You can "bust" downward, at worst trading two units for two units 2011-10-25T23:32:16 so yeh , you need to have a cost mechanic on the surrounding tiles 2011-10-25T23:32:19 Hey, what would happen if you were to run a thread in the background? 2011-10-25T23:32:24 influence map, maybe :) 2011-10-25T23:32:32 you'd be disqualified 2011-10-25T23:32:37 Is that against the rules, or would they just suspend your process after your turn? 2011-10-25T23:32:38 you'd be a very very naughty boy 2011-10-25T23:32:49 and yeh, it's against the rules 2011-10-25T23:32:49 burny, I know a process is banned, but is a thread? 2011-10-25T23:32:50 they send SIGSTOP to your program after each turn 2011-10-25T23:32:55 but you're not allowed to use more than 1 core 2011-10-25T23:32:55 but if you have a wider area than the enemy, they can never bust through 2011-10-25T23:33:11 Through that particular choke point, yep 2011-10-25T23:33:18 actually no 2011-10-25T23:33:20 i disgagree 2011-10-25T23:33:23 *** svujic has quit IRC (Quit: leaving) 2011-10-25T23:33:24 rmmh, green threads? 2011-10-25T23:33:29 XooooX 2011-10-25T23:33:32 *** Larose has joined #aichallenge 2011-10-25T23:33:33 XXooXX 2011-10-25T23:33:40 But yeah, if they send SIGSTOP, that makes it impossible. 2011-10-25T23:33:40 when you move your two ants into that choke point 2011-10-25T23:33:51 the two ants in the middle on the opposing side will also die, no? 2011-10-25T23:33:55 no 2011-10-25T23:33:55 Well, unless you were running a c program and worked some magic or something 2011-10-25T23:34:04 pipeep: coroutines are fine, if you don't use more than 1 physical core and you stop when they tell you to 2011-10-25T23:34:05 because there will be ants behind the 2 ants 2011-10-25T23:34:20 so it'll be 2 vs 2+ ants 2011-10-25T23:34:21 i should make some test maps.. 2011-10-25T23:34:25 ok 2011-10-25T23:34:28 look at this 2011-10-25T23:34:29 Actually, it depends on the attack range... 2011-10-25T23:34:32 http://i.imgur.com/taFkt.png 2011-10-25T23:34:56 yeh if you move ants towards the tips of the arc 2011-10-25T23:35:05 When you draw the areas an ant can attack, it either makes a point or it makes a flat surface... 2011-10-25T23:35:10 there will be exchanges there i think 2011-10-25T23:35:27 So the range affects whether you will exchange at the choke or whether one side will purely lose 2011-10-25T23:35:34 i got to read the battle mechanics again 2011-10-25T23:35:34 yes 2011-10-25T23:35:47 hmm 2011-10-25T23:35:51 E.g. attackRange2 = 9 versus attackRange2 = 8 2011-10-25T23:35:56 *** svujic has joined #aichallenge 2011-10-25T23:36:06 *** Slater has quit IRC (Ping timeout: 265 seconds) 2011-10-25T23:36:14 Question 2011-10-25T23:36:29 will your bot have internet access? 2011-10-25T23:36:50 I think you're not allowed to do what you want to do 2011-10-25T23:36:56 You could just use a cloud service to process the turns 2011-10-25T23:37:19 Exception, I don't think you can even make network calls with your bot, can you? 2011-10-25T23:37:29 didn't try 2011-10-25T23:37:31 *** djr__ has joined #aichallenge 2011-10-25T23:37:38 don't wanna be disqualified 2011-10-25T23:37:38 i bet this won't work but.. 2011-10-25T23:37:40 @learn tcp as http://ants.fluxid.pl/howto 2011-10-25T23:37:40 Alexer: Ready to serve. 2011-10-25T23:37:48 for real :o 2011-10-25T23:37:51 @tcp 2011-10-25T23:37:52 Alexer: tcp could be (#1) http://ants.fluxid.pl/, or (#2) http://ants.fluxid.pl/howto. 2011-10-25T23:37:53 I didn't try either, but I figured that it would be obvious 2011-10-25T23:38:04 *** chiller- has joined #aichallenge 2011-10-25T23:38:18 ...close but no cigar :( 2011-10-25T23:38:40 So each bot can use 1 GB of RAM 2011-10-25T23:38:56 their servers must have a lot of RAM lol 2011-10-25T23:39:18 *** SqrT_ has quit IRC (Ping timeout: 252 seconds) 2011-10-25T23:39:35 *** staalebk has quit IRC (Read error: Operation timed out) 2011-10-25T23:39:46 @list 2011-10-25T23:39:46 Alexer: AIChallenge, Admin, Alias, Anonymous, BadWords, Channel, ChannelLogger, ChannelStats, Config, Ctcp, Dict, Dunno, Factoids, Filter, Format, Games, Google, Herald, Internet, Later, Limiter, Math, Misc, Network, News, Note, Owner, Plugin, Quote, QuoteGrabs, RSS, Reply, Scheduler, Seen, Services, ShrinkUrl, Status, String, Success, Time, Todo, Topic, URL, Unix, User, Utilities, and Web 2011-10-25T23:39:58 *** JamesMG_ has quit IRC (Ping timeout: 252 seconds) 2011-10-25T23:40:11 *** JamesMG has joined #aichallenge 2011-10-25T23:40:35 I find AI to be really cool 2011-10-25T23:41:04 *** djr_ has quit IRC (Ping timeout: 252 seconds) 2011-10-25T23:41:45 < Exception> didn't try 2011-10-25T23:41:51 i did, though :D 2011-10-25T23:42:04 accidentally left in my logging (using sockets) 2011-10-25T23:42:22 i spent forever wondering why my bot was timing out 2011-10-25T23:43:41 turns out trying to connect to a random port on localhost is likely to cause timeouts.. >.>' 2011-10-25T23:43:49 *** svujic has left #aichallenge 2011-10-25T23:44:58 lol 2011-10-25T23:45:52 amstan, sorry, i broke your @tcp :( 2011-10-25T23:46:09 @forget tcp 2 2011-10-25T23:46:09 amstan: I come to serve. 2011-10-25T23:47:03 amstan, if you make it http://ants.fluxid.pl/howto they'll know what to do too, though 2011-10-25T23:47:12 @forget tcp 1 2011-10-25T23:47:12 amstan: As you wish. 2011-10-25T23:47:19 @learn tcp as http://ants.fluxid.pl/howto 2011-10-25T23:47:19 amstan: Ready to serve, my lord. 2011-10-25T23:47:23 *** Craklyn has quit IRC (Quit: Page closed) 2011-10-25T23:47:56 * Alexer hides 2011-10-25T23:48:26 is it against the rules to use the RF noise of the cpu, to send signals to a remote host, and use an external gamma ray laser, to twiddle bits in memory? 2011-10-25T23:48:41 burny: do so! 2011-10-25T23:49:13 * burny is ammused, trying to imagine someone attempting to debug said code 2011-10-25T23:53:00 burny, you should use butterflies: http://xkcd.com/378/ :) 2011-10-25T23:54:21 lol :) 2011-10-25T23:54:49 *** GGG has quit IRC (Quit: Page closed) 2011-10-25T23:58:34 *** analyst74 has joined #aichallenge