2011-11-02T00:00:08 non existent, kaput 2011-11-02T00:00:29 *** dancole has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 7.0.1/20110930134807]) 2011-11-02T00:00:32 ccc_: what is your username on the site? 2011-11-02T00:00:41 contestbot: seen j3camero 2011-11-02T00:00:41 amstan: j3camero was last seen in #aichallenge 1 week, 3 days, 7 hours, 47 minutes, and 5 seconds ago: If you are finding that some of the testing tools suck, please consider making better ones. We generally release whatever tools at the beginning, but the contestants always end up making better ones. 2011-11-02T00:01:05 McLeopold: username: rajasharan, userId 75 2011-11-02T00:01:13 amstan: for the love of ants, please split the forums up! 2011-11-02T00:01:42 McLeopold: i would like to go to sleep, how about tomorrow? or i can make you an admin 2011-11-02T00:01:51 *** rajanaresh has joined #aichallenge 2011-11-02T00:02:23 McLeopold: nvm, you are an admin 2011-11-02T00:03:25 @amstan: in many of the games my bot just crashed, and it didn't say as timeout, and i'm really surprised as to what made the bot crashed, sometimes it crashed in the middle of the game, sometimes at turn 1 itself, 2011-11-02T00:03:26 ccc_: I'm sorry Dave, err ccc_; I cannot 'amstan:'. 2011-11-02T00:03:36 McLeopold: i dont uundertand, i have to put " fprintf(STDERR, "hello"); " in my code and upload it ? 2011-11-02T00:03:41 aichallenge: McLeopold epsilon * r508b78c / website/visualizer.php : added admin view to errors - http://git.io/JeQqlg 2011-11-02T00:04:01 ccc, it's doing something wrong 2011-11-02T00:04:10 ccc_: you have some invalid moves, but that's not it. 2011-11-02T00:04:39 ccc_: can you do a catch all to exceptions and print a stack trace or something? 2011-11-02T00:05:24 capa: are you familiar with the code that is outputing the orders? 2011-11-02T00:05:34 hmm, my bot is just a simple C program, so i don't know if i can catch exceptions, also in my local it never crashes 2011-11-02T00:05:40 not so much 2011-11-02T00:07:37 ccc_: maybe you can setup a vm running ubuntu and try and make it crash 2011-11-02T00:07:53 capa: are you familiar with fprintf? 2011-11-02T00:08:36 McLeopold: i only know how to use stdout 2011-11-02T00:08:47 dont know stderr 2011-11-02T00:09:13 it's almost the same for console apps, it goes to the screen 2011-11-02T00:09:16 McLeopold: from the results I'm seeing I don't see how the seq count for opponent list used for the pareto cutoff can actually be counting in order of descending match quality. 2011-11-02T00:09:43 janzert: can you just pretend that it does :) 2011-11-02T00:10:06 it's my current theory on what is messing up pairings :P 2011-11-02T00:10:15 capa: fprintf(stderr, "blah"); will be ignored, but we will capture the output and show you if you crash 2011-11-02T00:10:26 janzert: how are pairings messed up? 2011-11-02T00:10:41 e.g. a group size of 11 with the closest player ranked over 150 ranks away can't be right 2011-11-02T00:11:01 they are pulling from too broad of rankings too consitently 2011-11-02T00:11:06 consistently 2011-11-02T00:11:10 my last match was #4, me, #4086, #935, #6 2011-11-02T00:11:15 void move(int index, char dir, struct game_state* Game, struct game_info* Info) { //fprintf(stdout, "O %i %i %c\n", Game->my_ants[index].row, Game->my_ants[index].col, dir); fprintf(stderr, "blah"); 2011-11-02T00:11:22 capa: so, just sprinkle some fprintf(stderr, ""); stuff and various points in your code, and if it crashes, all the prints below that point won't get printed, thus telling you where it possibly crashed 2011-11-02T00:11:56 janzert: for multiple opponents, a new pareto number is chosen for each one 2011-11-02T00:12:13 so the likelyhood of a tight range goes down 2011-11-02T00:12:31 I could do the pareto once, and use it for all opponents 2011-11-02T00:12:32 yes, the group I was talking about was the selection for one opponent (the first opponent) 2011-11-02T00:12:39 ok 2011-11-02T00:12:42 I'm looking at call opponent 2011-11-02T00:12:46 results 2011-11-02T00:12:48 *** dr- has quit IRC (Ping timeout: 240 seconds) 2011-11-02T00:12:58 there are nested queries doing the work 2011-11-02T00:13:10 the inside gets all submissions and ranks by match quality 2011-11-02T00:13:18 yeah, I know and reading the code seems like it should work :P 2011-11-02T00:13:27 it's just not what I'm seeing happen though 2011-11-02T00:13:31 the next layer increments the seq by one and attaches it to the results 2011-11-02T00:13:32 capa, as McLeopold said, printf("hello") is the same as fprintf(stdout, "hello") .... use stdout for moving ants, use stderr to output anything you want to your profile for debugging stuff 2011-11-02T00:14:01 janzert: can you pastebin a bad result? 2011-11-02T00:14:02 seq is the same for all the inside layer results? 2011-11-02T00:14:41 http://pastebin.com/KL6Pfhnd is the most current one I'm looking at 2011-11-02T00:15:14 McLeopold: do you think if I don't free the memory i malloc'd, the bot could crash 2011-11-02T00:15:58 *** dvladim has quit IRC (Ping timeout: 256 seconds) 2011-11-02T00:16:34 *** Antimony has joined #aichallenge 2011-11-02T00:17:32 can i get root access to run "ln -sf /usr/bin/tclsh8.5 /usr/bin/tclsh" on each server? I won't hack anything, I swear! 2011-11-02T00:17:36 ccc_: if you allocate over 1.5GB it will 2011-11-02T00:18:14 paulwal: :) no 2011-11-02T00:18:18 heh 2011-11-02T00:18:39 besides if you just logged in and did that it wouldn't help you any anyway 2011-11-02T00:18:49 since it wouldn't effect the sandbox at all ;) 2011-11-02T00:18:59 janzert: match quality looks like it counts down, and the result count is correct according to the random pareto number... 2011-11-02T00:19:09 so, the only problem could be missing entries 2011-11-02T00:19:22 it wouldn't? 2011-11-02T00:19:23 janzert: no I'm not allocating that much anyway, also one of the game crashed at just step 1, I took the "Bot In" results from the replay and fed it to my program, my bot worked fine 2011-11-02T00:19:26 what does that command even do? All I can tell is that it creates a symbolic link to tclsh 2011-11-02T00:19:50 yes, I just can't believe that there isn't a higher match quality within the closer 200+ ranks 2011-11-02T00:20:06 McLeopold: I can manually look for a few if you want 2011-11-02T00:20:07 the ranks that are outside the norm have high sigma 2011-11-02T00:20:52 janzert: I can create a query that gives the full match quality for the closest in rank 2011-11-02T00:20:54 McLeopold: I would expect all of them to be close to 220 they are all 306 or less 2011-11-02T00:20:59 Antimony, /usr/bin/tclsh is a symbolic link to /usr/bin/tclsh8.4. This would overwrite tclsh to point to /usr/bin/tclsh8.5. The compiler.py executes tclsh 2011-11-02T00:21:10 paulwal: thanks, i think it shows me where it crashed 2011-11-02T00:21:26 none higher is also a little suspicious 2011-11-02T00:21:31 higher than 220 that is 2011-11-02T00:21:41 yeah 2011-11-02T00:21:46 Antimoney, the long term fix is to have compiler.py execute /usr/bin/tclsh8.5 instead of tclsh 2011-11-02T00:21:51 I would expect it to be different as well 2011-11-02T00:21:53 Antimony* 2011-11-02T00:22:10 maybe the exp(sum(ln())) on a single result is causing rounding errors? 2011-11-02T00:22:19 no... 2011-11-02T00:22:25 so you're basically asking them to update the version of tclsh they use? 2011-11-02T00:22:32 second sig fig varies too much 2011-11-02T00:22:49 McLeopold: Can i use "malloc" ? 2011-11-02T00:22:57 yeah, did you see my query asking why you don't just do a real geometric mean? 2011-11-02T00:22:57 sure 2011-11-02T00:23:07 no, I missed that 2011-11-02T00:23:39 dividing the sum result by (@player_count + 1) before get the exp of it would be a real geometric mean 2011-11-02T00:23:48 just wondering if there was a reason you didn't? 2011-11-02T00:23:54 nope 2011-11-02T00:23:56 it doesn't make a difference in ordering of course 2011-11-02T00:24:07 just more intuitive to look at I suppose 2011-11-02T00:24:16 Antimony, yes. tclsh8.5 is already installed. The package is installed in new servers by setup/worker_setup.py. But worker/compiler.py executes tclsh, which by default is a symlink to /usr/bin/tclsh8.4 instead of /usr/bin/tclsh8.5 2011-11-02T00:24:48 wow, trueskill is implemented as a stored procedure 2011-11-02T00:25:04 a1k0n: match quality only, just an approximation 2011-11-02T00:25:06 a1k0n: just the pairing not rank updating 2011-11-02T00:25:18 ah i see 2011-11-02T00:26:39 janzert: another crash for skill points: http://aichallenge.org/profile.php?user=389 2011-11-02T00:27:01 yeah, I saw it already. the mu is consitently going down of course 2011-11-02T00:27:02 Antimony, so I'm thinking the `ln` command above would be an immediate fix (because i'm being impatient), but janzert has informed me that the sandbox would not see this change. 2011-11-02T00:27:08 just sigma is moving faster still 2011-11-02T00:27:52 paulwal: i'm guessing you can't just put #!/usr/bin/tclsh8.4 at the top? 2011-11-02T00:28:17 er, 8.5 2011-11-02T00:28:35 McLeopold: I did what you said, and i think it shows me where it crashed. but i still dont know what is happening .. 2011-11-02T00:28:43 a1k0n: no the worker is explicitly running it with tclsh 2011-11-02T00:28:59 a1k0n, i don't think it would work 2011-11-02T00:29:59 tclsh would have to be aware of the various versions, etc, like perl 2011-11-02T00:31:01 * janzert plans on starting the worker update in about 15 minutes 2011-11-02T00:31:17 * paulwal thanks janzert kindly :) 2011-11-02T00:31:26 *** v_paul_v_ has joined #aichallenge 2011-11-02T00:31:26 besides sampling from a pareto distribution, isn't there value in ensuring a "bracketing" for the selected player? 2011-11-02T00:32:03 a1k0n, i'm pretty certain tclsh just ignores the hashbang as a comment 2011-11-02T00:32:13 a1k0n: match_quality should be taking care of it 2011-11-02T00:32:22 e.g. i just got paired up with two players ranked well below me, and i won, which reduces my sigma but doesn't add much information 2011-11-02T00:32:38 yeah, something seems broken atm 2011-11-02T00:33:33 janzert: http://pastebin.com/ekjHMH5x 2011-11-02T00:34:19 a1k0n: if you get a large pareto number, it will favor a round-robin approach with the nearer bots first 2011-11-02T00:34:53 so if the random number was very large (1000), all 999 bots would have to have played a game before 1000 is actually picked 2011-11-02T00:35:04 McLeopold: that looks pretty good 2011-11-02T00:35:15 brb 2011-11-02T00:35:20 *** paulwal has quit IRC (Quit: Page closed) 2011-11-02T00:35:38 janzert: notice how the top ranked in match quality are actually far away in rank? 2011-11-02T00:35:38 *** tsetliff has joined #aichallenge 2011-11-02T00:35:41 yeah what the hell? that sum only had one argument 2011-11-02T00:35:59 way closer than what that last one was 2011-11-02T00:36:00 exp(sum(ln(...))) 2011-11-02T00:36:16 McLeopold: OH, i got "6 Nov 2nd 4:34am Success: ready to play" 2011-11-02T00:36:25 ty man 2011-11-02T00:36:55 and better match qualities than any of the opponent call results 2011-11-02T00:37:21 or they're not directly comparable I guess 2011-11-02T00:37:35 well, they should be 2011-11-02T00:38:28 is beta 1/sigma? 2011-11-02T00:38:50 in your terminology there 2011-11-02T00:38:56 beta = initial mu / 6 2011-11-02T00:39:02 http://pastebin.com/6EhvJyWG 2011-11-02T00:39:04 where initial mu = 50 2011-11-02T00:39:09 well yes, i see that 2011-11-02T00:39:11 but what does beta mean 2011-11-02T00:39:24 janzert: the exp(sum(ln())) does not change the results for a single opponent 2011-11-02T00:39:43 it's basically a measure of the variance in game results if I recall correctly 2011-11-02T00:40:01 oh. so it's from the canonical trueskill equations 2011-11-02T00:40:05 lemme pull that up 2011-11-02T00:40:07 McLeopold: right I realized that after a second or two ;) 2011-11-02T00:40:16 janzert, McLeopold I'm not that familiar with how TrueSkill works under the covers (I was just skimming the MS site for it), and maybe this is totally irrelevant, but could the funny matches be caused by the system trying to keep the game rate even and thus eliminating a lot of bots from contention because they have played too many other games recently? 2011-11-02T00:40:16 janzert: okay, gtg 2011-11-02T00:40:44 ChrisH_: yes, that is part of it 2011-11-02T00:40:46 *** djr_ has quit IRC (Read error: Connection reset by peer) 2011-11-02T00:40:50 McLeopold: i simulated my game in ubuntu just like you said and it seg-faulted, but in my Mac it runs just fine, so anyway thanks a ton 2011-11-02T00:41:12 *** djr_ has joined #aichallenge 2011-11-02T00:43:32 McLeopold: so how are these better pairings being dropped for the 'call opponent' results? 2011-11-02T00:43:43 anyway I'll try and dig into it myself 2011-11-02T00:45:08 ah, beta is the variance of performance. which should be pretty low (if not 0) for a bot 2011-11-02T00:46:17 heh but all kinds of math breaks down at 0. makes sense. 2011-11-02T00:46:56 a1k0n: I believe it's the variance of performance as in how likely a truly lower ranked player is to beat a higher ranked player not variance for one player between different games 2011-11-02T00:47:12 but then I haven't looked at that in quite a while so I may be remembering wrong 2011-11-02T00:47:20 well, it works out the same 2011-11-02T00:47:25 not quite 2011-11-02T00:47:42 beta for ants is probably still fairly high even with completely static players 2011-11-02T00:47:58 agreed 2011-11-02T00:48:02 *** zigner22 has joined #aichallenge 2011-11-02T00:48:20 the implications may be different but the math is the same 2011-11-02T00:49:31 ok yes :) just we shouldn't set beta extremely low because the bots performance themselves shouldn't vary 2011-11-02T00:49:45 right, yes. 2011-11-02T00:50:19 your spawn location can get occasionally scouted randomly on some maps and there's nothing you can do about it 2011-11-02T00:50:32 etc etc 2011-11-02T00:54:50 *** tsetliff has quit IRC (Quit: Page closed) 2011-11-02T00:54:52 the paper alludes to a dynamics variance separate from a performance variance 2011-11-02T00:55:03 but it doesn't seem to.. use it 2011-11-02T00:58:22 oh that's change in skill over time 2011-11-02T01:02:32 *** ccc_ has quit IRC (Ping timeout: 265 seconds) 2011-11-02T01:02:55 *** paulwal has joined #aichallenge 2011-11-02T01:05:31 hey all, my profile has been saying "Next game should be within 134.7 minutes." for the while evening 2011-11-02T01:05:40 is that normal? 2011-11-02T01:06:18 it played a game at 1:49am and now its 5am 2011-11-02T01:06:31 treeform: I wouldn't rely on the official servers giving you games regularly due to all the cpu limiting 2011-11-02T01:06:38 not for testing anyway 2011-11-02T01:07:08 is there other servers for testing? 2011-11-02T01:07:15 treeform: is the number in the "Next Game: #### players are ahead." part going down? 2011-11-02T01:07:25 treeform: http://ants.fluxid.pl/howto 2011-11-02T01:07:42 yes it is 2011-11-02T01:07:50 then you are getting closer to another game 2011-11-02T01:08:11 but you shouldn't expect more than a handful of games per day on the main server at the moment 2011-11-02T01:08:25 will that improve? 2011-11-02T01:08:32 unknown 2011-11-02T01:08:40 or is that how it will be till the challenge ends in december? 2011-11-02T01:09:03 I don't know, so much could change between now and then 2011-11-02T01:09:09 *** replore_ has quit IRC (Remote host closed the connection) 2011-11-02T01:09:10 *** kiv has joined #aichallenge 2011-11-02T01:09:20 and I'm just another player like you, so I don't know anything other than what I read here and on the forums 2011-11-02T01:09:39 hmm ok 2011-11-02T01:10:02 a1k0n: in the http://ants.fluxid.pl/howto username and pass? where would i get that? 2011-11-02T01:10:08 but I've been around for a while, so I do know that they want it to be better, but it isn't easy with the number of contestants vying for resrouces right now 2011-11-02T01:10:17 ok 2011-11-02T01:11:06 treeform: you make them up; accounts are implicitly created the first time 2011-11-02T01:11:14 treeform: but I can tell you that when the final competition is run they will likely prune the bottom of the standing away after some amount of time to dedicate more horsepower to getting the top ranks as accurate as possible, at least that's what they did in the last competition 2011-11-02T01:11:17 it just prevents impersonation once you've established a ranking 2011-11-02T01:11:51 ChirsH_ ok 2011-11-02T01:12:01 and got it on the pw 2011-11-02T01:12:03 *** timos has quit IRC (Ping timeout: 245 seconds) 2011-11-02T01:12:30 *** kiv has quit IRC (Client Quit) 2011-11-02T01:12:42 *** Knekkebjoern has joined #aichallenge 2011-11-02T01:12:47 *** timos has joined #aichallenge 2011-11-02T01:16:32 oh i get it. you're taking the product of all the draw probabilities 2011-11-02T01:17:25 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-11-02T01:17:26 *** choas has joined #aichallenge 2011-11-02T01:18:33 *** replore has joined #aichallenge 2011-11-02T01:18:55 *** heinrich5991_ has joined #aichallenge 2011-11-02T01:19:08 *** heinrich5991 has quit IRC (Ping timeout: 260 seconds) 2011-11-02T01:19:08 *** heinrich5991_ is now known as heinrich5991 2011-11-02T01:19:09 *** Tumi_ has quit IRC (Ping timeout: 260 seconds) 2011-11-02T01:19:25 *** Tumi_ has joined #aichallenge 2011-11-02T01:19:44 *** oddmunds has quit IRC (Ping timeout: 260 seconds) 2011-11-02T01:20:01 *** oddmunds has joined #aichallenge 2011-11-02T01:20:02 *** xar0l has joined #aichallenge 2011-11-02T01:24:59 *** ccc_ has joined #aichallenge 2011-11-02T01:25:11 *** Venom has joined #aichallenge 2011-11-02T01:25:22 capa: what kind of problem were you having with your code, i just solved mine 2011-11-02T01:25:45 *** rajanaresh has quit IRC (Ping timeout: 260 seconds) 2011-11-02T01:25:57 *** theskumar has joined #aichallenge 2011-11-02T01:28:03 Regarding , is our local game engine generates those (by default) very similarly like the contest server does? 2011-11-02T01:30:24 *** Areks_ has joined #aichallenge 2011-11-02T01:31:43 My costmap resists all attempts at modifying it, I am puzzled 2011-11-02T01:32:07 the distance() function in the c starter pack is wrong, just fyi everyone 2011-11-02T01:33:10 theres an easier way to do it 2011-11-02T01:38:22 <_flag> Which c starter pack are you using? 2011-11-02T01:38:49 *** ccc_ has quit IRC (Quit: Page closed) 2011-11-02T01:39:05 *** bmh has quit IRC (Quit: bmh) 2011-11-02T01:40:22 <_flag> bysin: Judging by the one listed on the website, that would appear to be my code, so I'm curious what you think I did wrong 2011-11-02T01:42:51 nice, a game between #1, #2, #9, and #3496 2011-11-02T01:44:59 *** stocha has joined #aichallenge 2011-11-02T01:45:08 @tcp 2011-11-02T01:45:09 stocha: tcp could be http://ants.fluxid.pl/howto. 2011-11-02T01:47:18 *** edgarymr has joined #aichallenge 2011-11-02T01:48:47 *** replore has quit IRC (Ping timeout: 252 seconds) 2011-11-02T01:50:43 *** stocha has quit IRC (Quit: Page closed) 2011-11-02T01:50:52 *** formallyknownasd has quit IRC (Ping timeout: 265 seconds) 2011-11-02T01:51:56 *** roflmao has left #aichallenge 2011-11-02T01:52:58 *** svujic has left #aichallenge 2011-11-02T01:54:12 a1k0n: how did the 3496 do? 2011-11-02T01:56:01 meh.. I gone back to my A* bot 2011-11-02T01:56:17 <_flag> a1k0n: xathis has been playing against a lot of very low players in recent games, I think it may be a bug in the new enable/disable matchmaking 2011-11-02T01:56:31 grom358, As opposed to? 2011-11-02T01:56:43 SharkMonkey: as opposed to BFS 2011-11-02T01:56:58 I use BFS for food collecting though 2011-11-02T01:57:18 _flag: janzert thinks there is a matchmaking problem too and was working on figuring out what's going on earlier 2011-11-02T01:57:44 *** Cyndre has quit IRC (Quit: Leaving) 2011-11-02T01:57:53 <_flag> Okay, as long as they know 2011-11-02T01:58:19 *** pairofdice has quit IRC (Read error: Operation timed out) 2011-11-02T01:58:20 *** Palmik has joined #aichallenge 2011-11-02T01:58:28 like the maps, it's wrapping at 0? :P 2011-11-02T01:58:44 <_flag> Also, does anyone get ridiculously high sigmas on tcp? I have 5000 games played and my rank isn't stable 2011-11-02T01:59:57 treeform: not well! 2011-11-02T02:01:06 There aren't many bots that do well agains xathis right now. 2011-11-02T02:01:30 what is xathis doing that is so good? 2011-11-02T02:02:16 winning 2011-11-02T02:02:23 good exploration, good hunting. shrug. 2011-11-02T02:02:32 most everything 2011-11-02T02:02:47 i need to implement combat 2011-11-02T02:02:50 <_flag> From the looks of it, he's doing everything the other top bots are doing but doing it better 2011-11-02T02:03:17 <_flag> Hopefully this won't be another case of bocimacko :) 2011-11-02T02:03:21 i need to implement combat too. 2011-11-02T02:03:43 <_flag> s/bocimacko/bocsimacko 2011-11-02T02:03:49 hello,all 2011-11-02T02:03:52 the top bots still have a LONG ways to go 2011-11-02T02:04:07 I think you're right burny 2011-11-02T02:04:31 i hope so, otherwise this contest ended before it began 2011-11-02T02:04:40 thanks to the long-ass beta phase 2011-11-02T02:04:53 <_flag> The other thing too is that it takes a really long time for ranks to stabilize on the main server 2011-11-02T02:05:03 no kidding 2011-11-02T02:05:11 <_flag> So if you submitted your bot early, you probably won't drop out of the top 100 for the rest of the contest 2011-11-02T02:05:25 lol 2011-11-02T02:05:33 <_flag> Especially with how volatile ants games can be 2011-11-02T02:05:46 that better be fixed then.. 2011-11-02T02:05:47 <_flag> To make things worse, nobody plays on tcp for long 2011-11-02T02:05:53 <_flag> So the player pool is really shallow 2011-11-02T02:05:56 cause if I start making my bot in a couple weeks.. 2011-11-02T02:06:00 and submit a week later 2011-11-02T02:06:11 and can't ever get a single game to test my bot on.. 2011-11-02T02:06:13 I'll be mad 2011-11-02T02:06:41 *** Fandekasp has joined #aichallenge 2011-11-02T02:07:25 *** pairofdice has joined #aichallenge 2011-11-02T02:14:15 Ahh, awesome, Python is screwing with my head but atleast I got 100% from #aiclass homework 2011-11-02T02:15:42 I have broken this into small pieces that all work separately and they should also work together but they don't 2011-11-02T02:16:54 pairofdice, you did the hw in python? 2011-11-02T02:17:09 Nono, that was on paper 2011-11-02T02:17:41 I'm trying to do BFS for my ant but I cant enter the costs into the map for some reason 2011-11-02T02:18:03 *** Knekkebjoern has quit IRC (Quit: Leaving.) 2011-11-02T02:21:37 Spooky 2011-11-02T02:24:32 *** ltriant has quit IRC (Quit: Computer has gone to sleep) 2011-11-02T02:24:41 http://aichallenge.org/visualizer.php?game=45543 at the end that felt a lot like a splatter film. 2011-11-02T02:27:49 *** ChrisH_ has quit IRC (Quit: Page closed) 2011-11-02T02:29:12 *** choas has quit IRC (Ping timeout: 240 seconds) 2011-11-02T02:29:59 why would bot be deactivated 2011-11-02T02:30:05 aka "not in queue to play a game"? 2011-11-02T02:31:07 gud morning everyone.! :) 2011-11-02T02:31:15 ruser: http://aichallenge.org/forums/viewtopic.php?f=3&t=1710 2011-11-02T02:32:17 bqf: thanks, 2011-11-02T02:32:38 *** cyphase has quit IRC (Ping timeout: 260 seconds) 2011-11-02T02:34:07 *** master_ninja has joined #aichallenge 2011-11-02T02:36:49 *** SharkMonkey has quit IRC (Remote host closed the connection) 2011-11-02T02:38:12 *** theskumar has quit IRC (Ping timeout: 256 seconds) 2011-11-02T02:40:39 *** NefariousZhen has quit IRC (Ping timeout: 265 seconds) 2011-11-02T02:41:30 *** Palmik has quit IRC (Remote host closed the connection) 2011-11-02T02:43:50 *** scribble has quit IRC () 2011-11-02T02:47:19 *** grom358 has quit IRC (Quit: Ex-Chat) 2011-11-02T02:51:18 moin 2011-11-02T03:01:26 *** xar0l has quit IRC (Ping timeout: 265 seconds) 2011-11-02T03:03:33 *** Fandekasp has quit IRC (Ping timeout: 255 seconds) 2011-11-02T03:07:49 *** Saulzar has quit IRC (Ping timeout: 240 seconds) 2011-11-02T03:09:45 *** theskumar has joined #aichallenge 2011-11-02T03:10:35 *** djr_ has quit IRC (Quit: Leaving) 2011-11-02T03:12:50 alright, I started on "combat" 2011-11-02T03:12:54 *** Overburn has joined #aichallenge 2011-11-02T03:13:05 atm I'm just helping my ants control their suicidal tendencies :P 2011-11-02T03:13:10 mornin everyone 2011-11-02T03:13:15 heh what do you mean thestinger? 2011-11-02T03:13:35 *** grom358 has joined #aichallenge 2011-11-02T03:13:39 wtf.. http://aichallenge.org/visualizer.php?game=44507 2011-11-02T03:13:43 if there are 2 enemy ants that can attack a square, I won't move there w/o 2 of my own 2011-11-02T03:13:46 it's a dumb heuristic basically 2011-11-02T03:13:54 it's never a bad thing, but it doesn't really help xD 2011-11-02T03:14:30 thestinger: how do you get them to move together? 2011-11-02T03:15:06 I have a queue of orders, my enemy avoidance stuff is basically just tacked on atm 2011-11-02T03:15:15 it will be a while before it really works 2011-11-02T03:15:30 I ended up dropping my reimplementation with BFS 2011-11-02T03:15:41 only using BFS for certain things 2011-11-02T03:16:03 I'm thinking about weighting my distance maps based on the combat info 2011-11-02T03:16:19 instead of just sort of tacking it on to my step() function which is pretty stupid 2011-11-02T03:16:49 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Read error: Operation timed out) 2011-11-02T03:17:02 yeah.. my avoidance is crap too.. what I'm gonna do is path find towards a friend.. and the friend paths finds towards it 2011-11-02T03:17:16 then hunt down the chasing ant 2011-11-02T03:17:59 need to build in some way to pair movements 2011-11-02T03:18:10 like.. move only if my friend is moving 2011-11-02T03:18:10 *** amstan has quit IRC (Ping timeout: 260 seconds) 2011-11-02T03:18:50 actually.. what would be better is have an ant pack 2011-11-02T03:19:25 atm I calculate the number of enemy ants that can attack squares and the number of my ants that can attack squares - w/o considering ants moving the next turn 2011-11-02T03:19:37 I need to figure out a good way of using that data to weight my pathfinding though 2011-11-02T03:19:41 thestinger: yeah I got that too.. but not sure how to use it 2011-11-02T03:20:13 *** tmandry has quit IRC (Ping timeout: 244 seconds) 2011-11-02T03:20:46 yeah, I don't want my explorers to completely avoid enemy ants - I am fine with trading an ant for an ant because I need to do it to be able to see a bit further 2011-11-02T03:21:19 thestinger how did you implement your queue? 2011-11-02T03:21:22 one for every ant 2011-11-02T03:21:23 yeah.. my explorers completly avoid atm.. which means they could get a piece of food if they didn't run 2011-11-02T03:21:38 i need to make it trade ant for ant 2011-11-02T03:21:45 or one for everything 2011-11-02T03:21:52 AHHH! Finally I got it working. I spent like 9 hours on this simplest thing 2011-11-02T03:22:19 Overburn: it's really hackish atm lol, I have a separate queue/movement cycle for each set of goals 2011-11-02T03:23:03 I put an ant onto the queue with 1) the ant's location and 2) each direction that will bring it closer to a target 2011-11-02T03:23:06 hey my bot made it into the top 200 without combat 2011-11-02T03:23:21 yeah, I've made it into the top 100 three times now w/o combat :P 2011-11-02T03:23:33 haha congrat grom 2011-11-02T03:23:42 *** theskumar has left #aichallenge 2011-11-02T03:23:57 I think ChrisH actually made it to the top 10 without combat, iirc he was just avoiding enemy ants at first 2011-11-02T03:24:05 seems my attack hill at all costs works a lot of the time 2011-11-02T03:24:07 I'm sure he has combat now tho 2011-11-02T03:24:52 I made one that did the top 400 without even considering any data about enemies (except for hills). 2011-11-02T03:25:16 yeah, my current code on the site doesn't deal with enemies at all 2011-11-02T03:25:29 locally I have the "combat" stuff started 2011-11-02T03:25:58 how can u do top 10 without combat.. cause u have to raze.. which means combat 2011-11-02T03:26:07 well, I'm not including that as combat 2011-11-02T03:26:27 yeah, i'm #40 and climbing and i just ignore enemy ants, and send everything after known enemy hills to raze them 2011-11-02T03:26:36 any one using nodejs and having problems with flushing? 2011-11-02T03:26:39 a1k0n: that is what i do 2011-11-02T03:26:49 grom: yes, it does if your "swarming" is good. sometimes it's just emerge and you don't even need swarming code 2011-11-02T03:26:56 yeah I explore when I don't know where an enemy hill is, then rush it 2011-11-02T03:27:10 yeah i've had entire armies wiped out by two ants stationed on either side of the enemy hill 2011-11-02T03:27:15 I worked on avoiding forming straight lines a bit, so they aren't completely helpless 2011-11-02T03:27:36 a1kon: haha. happenned to me too. 2011-11-02T03:27:44 yeah that happens to me 2011-11-02T03:27:48 I don't have proper foraging atm because I don't re-explore locations - I just haven't gotten around to finishing up that code 2011-11-02T03:27:55 2 ants take out 30 - 50 of mine 2011-11-02T03:28:11 thestinger is in the top10? 2011-11-02T03:29:06 no 2011-11-02T03:29:07 ChrisH 2011-11-02T03:29:18 ah, okay 2011-11-02T03:29:26 anyway, I'm sure he has combat now 2011-11-02T03:29:31 *** delt0r__ has quit IRC (Ping timeout: 244 seconds) 2011-11-02T03:29:51 but he made it there by posting sentinels around his hills and avoiding enemies when exploring/gathering if he wasn't hiding something :P 2011-11-02T03:30:54 btw, if the AI limit is ultimately the processing time, how are some languages supposed to compete with some others *way* faster? 2011-11-02T03:31:24 you've just discovered the rationale for my using C++ 2011-11-02T03:31:40 a1k0n: please don't kill me; I won't tell 2011-11-02T03:31:48 :) 2011-11-02T03:31:52 yeah, I learned C++ for this 2011-11-02T03:31:55 i'm actually super amazed with GarySWest doing so well in python 2011-11-02T03:32:06 I already knew C really well, so it wasn't really a big deal to learn it 2011-11-02T03:32:22 *** TheDigitalNinja has quit IRC (Remote host closed the connection) 2011-11-02T03:32:55 a1k0n: I bet he's using numpy/scipy 2011-11-02T03:33:06 I'm in C#, so that wouldn't be as far as Ruby/Python->C++ .. but I hope I won't need to go there... 2011-11-02T03:33:12 mmm. perhaps. 2011-11-02T03:33:23 I don't see why else he would be using py2 and not py3 2011-11-02T03:33:36 *** balkemacho has joined #aichallenge 2011-11-02T03:35:13 if the server had pypy I might still be using python - 10x slower than C is much easier to deal with than 60x slower 2011-11-02T03:36:03 Notice for the finals: Turn time will be adjusted for each language to balance them. The balancing are based on benchmark data. Good Luck! :P 2011-11-02T03:36:40 *** treeform has quit IRC (Quit: Leaving.) 2011-11-02T03:36:56 interesting. 2011-11-02T03:37:09 don't be misled ;) 2011-11-02T03:37:30 yeah can i have a 60-second turn in python 2 while i use numpy? 2011-11-02T03:37:36 oops.. here's the tag ;) 2011-11-02T03:37:45 *** Jak_o_Shadows1 has joined #aichallenge 2011-11-02T03:37:47 *** onensora has joined #aichallenge 2011-11-02T03:38:02 *** Cyb has joined #aichallenge 2011-11-02T03:38:04 yeah, you can upload a python script that just execs compiled code 2011-11-02T03:38:06 Venom: fair is a level playing field with all bots receiving the same resources, how you use those resources is up to you :) 2011-11-02T03:38:55 *** Jak_o_Shadows has quit IRC (Ping timeout: 252 seconds) 2011-11-02T03:40:56 *** Jak_o_Shadows1 is now known as Jak_o_Shadows 2011-11-02T03:40:56 janzert: I agree. Would you be so kind as to accept my X86 Assembly code? Also, any juicy GPU in the server? I'd love to get OpenCL-fair. 2011-11-02T03:41:07 ;) 2011-11-02T03:41:22 nope GPU carrying servers are quite expensive :) 2011-11-02T03:41:42 x86 assembly is very possible if you want to add the support :) 2011-11-02T03:41:48 http://www.ibiblio.org/gferg/ldp/GCC-Inline-Assembly-HOWTO.html :P 2011-11-02T03:41:50 *** Saulzar has joined #aichallenge 2011-11-02T03:42:25 *** delt0r__ has joined #aichallenge 2011-11-02T03:43:41 janzert: that'd be cool, but I'm not crazy (in a good way) enough. 2011-11-02T03:43:50 hehe 2011-11-02T03:44:22 yeah, I'd be impressed by an assembly bot but more scared of its author ;) 2011-11-02T03:44:27 gcc has a lot of built-in functions starting with __ for doing SIMD stuff 2011-11-02T03:44:58 yep gcc intrinsics gave me some nice speedups for my tron bot 2011-11-02T03:47:57 *** Saulzar has quit IRC (Ping timeout: 258 seconds) 2011-11-02T03:49:22 I guess I can just store last_seen = turn for each location and revisit them after a while for food gathering 2011-11-02T03:53:32 *** userjjb has joined #aichallenge 2011-11-02T03:55:02 noob question, I downloaded the tools package in WinXP and installed Python 2.7.2, but when I run "play_one_game.cmd" the command window flashes up and then closes itself immediately 2011-11-02T03:55:27 *** Cyb has quit IRC (Ping timeout: 248 seconds) 2011-11-02T03:55:50 *** zigner22 has quit IRC (Ping timeout: 256 seconds) 2011-11-02T03:56:18 read the tutorial and make the tutorial.cmd 2011-11-02T03:57:35 make the tutorial in cmd 2011-11-02T03:57:40 *** ^5 <^5!7cb29512@gateway/web/freenode/ip.124.178.149.18> has joined #aichallenge 2011-11-02T03:57:43 and python path into win 2011-11-02T03:58:05 *** caution has joined #aichallenge 2011-11-02T03:58:20 I keep playing games where most of the players timeout, including me 2011-11-02T03:58:26 is it normal? 2011-11-02T03:59:30 i donot know,i just begin 2011-11-02T04:00:48 *** master_ninja has quit IRC (Quit: master_ninja) 2011-11-02T04:02:09 *** avdg1 has joined #aichallenge 2011-11-02T04:02:57 what's "skill" a measure of? 2011-11-02T04:04:23 *** FastEddie_ has joined #aichallenge 2011-11-02T04:04:35 *** avdg has quit IRC (Ping timeout: 252 seconds) 2011-11-02T04:05:22 *** Garf has joined #aichallenge 2011-11-02T04:06:40 i do not know 2011-11-02T04:08:25 Arf, wall of duplicate order critical hit 2011-11-02T04:09:56 ok fixed my first problem, just turned out I needed to define a path variable for python 2011-11-02T04:10:07 surprised it doesn't do that on install 2011-11-02T04:10:26 Yeah, it's a bit derp 2011-11-02T04:11:25 In the visualizer that shows the battle replay is there a way to turn off the display of the wrapped level on all sides of the actual map? 2011-11-02T04:14:55 *** SharkMonkey has joined #aichallenge 2011-11-02T04:19:38 *** HaraKiri has joined #aichallenge 2011-11-02T04:19:55 *** mleise has joined #aichallenge 2011-11-02T04:20:04 *** Chunes has joined #aichallenge 2011-11-02T04:23:22 *** mleise has quit IRC (Client Quit) 2011-11-02T04:24:32 Runs fine locally but gets index error on server... errwhat 2011-11-02T04:26:19 are you testing on different maps? 2011-11-02T04:27:11 *** cjoe has joined #aichallenge 2011-11-02T04:28:21 Ahh, okay, timeout 2011-11-02T04:28:41 *** epicmonkey has joined #aichallenge 2011-11-02T04:32:30 *** Blkt has joined #aichallenge 2011-11-02T04:33:23 *** Kingpin13 has quit IRC (Quit: quit) 2011-11-02T04:33:37 *** mviel_ has joined #aichallenge 2011-11-02T04:36:39 *** Chunes has quit IRC (Ping timeout: 265 seconds) 2011-11-02T04:38:07 *** epicmonkey has quit IRC (Ping timeout: 248 seconds) 2011-11-02T04:38:20 *** Jak_o_Shadows1 has joined #aichallenge 2011-11-02T04:41:20 *** ikaros has joined #aichallenge 2011-11-02T04:41:37 *** Jak_o_Shadows has quit IRC (Ping timeout: 258 seconds) 2011-11-02T04:42:11 If food appears next to an ant, does it need to stay there for a turn to consume it? 2011-11-02T04:42:36 yes 2011-11-02T04:42:36 yes, you can actually get blocked if you try to move onto a food 2011-11-02T04:44:16 And dead when an ant behind you moves in because it assumed you were stepping out. :( 2011-11-02T04:44:51 just check for food when you step 2011-11-02T04:44:58 * SharkMonkey is way too tired to reason about his program anymore 2011-11-02T04:45:22 Same 2011-11-02T04:45:47 for each food, I send the nearest ant (well, not exactly b/c I'm working on code to reuse ants), so if that ant is 1 square away from the food I tell it not to move that turn 2011-11-02T04:46:04 and I prevent ants from stepping on food just like water now, to avoid the same collision thing you just said 2011-11-02T04:46:14 *** Bluedgis has joined #aichallenge 2011-11-02T04:46:47 I actually noticed it b/c I keep seeing "move blocked" or w/e 2011-11-02T04:46:53 and I realized it was b/c of food 2011-11-02T04:47:36 alright, re-exploring all squares after they haven't been seen for 30 turns ruins my regular exploration :P 2011-11-02T04:48:49 @later tell McLeopold pretty sure I found the pairing problem, line 215 of 2_generate_matchup.sql is actively fighting against the deactivation mechanism and in the mean time causing worse and worse matchups. Going to just comment it out for now as I think we'll probably be fine without it completely, but we can probably come up with something that doesn't fight deactivation if needed 2011-11-02T04:48:49 janzert: Yes master! 2011-11-02T04:52:27 aichallenge: janzert epsilon * rd1ab412 / (sql/2_generate_matchup.sql sql/opponent.sql): Stop pairing from fighting against deactivation - http://git.io/otcJIQ 2011-11-02T04:53:44 Well, atleast I got it to work on one map... >_< 2011-11-02T04:54:36 *** mleise has joined #aichallenge 2011-11-02T04:56:18 @later tell McLeopold that change went live as of game 47650 btw 2011-11-02T04:56:18 janzert: Aye, aye, sir 2011-11-02T04:59:24 i made my first customization of the tutorial bot.. and now it crashes.. this was not how i imagined it to be! cue painful debugging with log files.. 2011-11-02T05:00:24 *** sigh has joined #aichallenge 2011-11-02T05:01:07 I guess I need to divide the map in to "rooms" for this foraging stuff >.< 2011-11-02T05:02:12 *** pairofdice has quit IRC (Quit: Just keep breathing) 2011-11-02T05:05:18 can you test timeouts with the local game? 2011-11-02T05:07:30 yes 2011-11-02T05:07:31 answer: yes 2011-11-02T05:07:42 set turntime to 80, use the 8 hill multi maze map 2011-11-02T05:08:18 I think EC2 can be much slower than just desktop/laptop hardware 2011-11-02T05:09:37 *** Minthos has quit IRC (Remote host closed the connection) 2011-11-02T05:09:40 *** xathis has joined #aichallenge 2011-11-02T05:09:54 *** Minthos has joined #aichallenge 2011-11-02T05:10:05 if one bot uses a lot of CPU does it slow the other bots? 2011-11-02T05:10:34 they each get a core iirc 2011-11-02T05:11:39 that's why you can't spawn processes/threads I guess - you would get an unfair time share from the kernel 2011-11-02T05:18:32 my crappy bot times out in do_setup() on that map 2011-11-02T05:18:35 pathetic 2011-11-02T05:18:58 all it does is build a dict of tiles 2011-11-02T05:19:26 hmm I did that with no problem in python 2011-11-02T05:19:28 you might have a bug 2011-11-02T05:20:25 also you could just build a 2d list instead of a dict (they are implemented as dynamic arrays so they have O(n) random access) 2011-11-02T05:20:52 and then do grid[row][col], instead of grid[(row, col)] 2011-11-02T05:21:17 caution: btw, you might find this useful http://wiki.python.org/moin/TimeComplexity 2011-11-02T05:21:32 python dicts are a hash table so with a decent hash they should give you O(1) access 2011-11-02T05:21:39 yeah 2011-11-02T05:21:47 I've speed tested them, dict is only fractionally slower 2011-11-02T05:22:01 of course the dict O(1) will still be larger than an array O(1) :) 2011-11-02T05:22:15 thestinger: why not use array.array? 2011-11-02T05:23:21 I guess that would be even better :P 2011-11-02T05:23:29 or use numpy arrays 2011-11-02T05:24:51 feels like you might as well use C++ if you have to optimise python so hard 2011-11-02T05:25:04 yeah, that's what I decided xD 2011-11-02T05:26:17 *** analyst74 has quit IRC (Ping timeout: 258 seconds) 2011-11-02T05:27:50 c++ <3 2011-11-02T05:28:09 *** mleise has quit IRC (Quit: Leaving.) 2011-11-02T05:29:04 yeah, you can use C++11 now 2011-11-02T05:29:16 so you get unordered_set and unordered_map :) 2011-11-02T05:30:00 python will probably actually be harder to use if you already know C or C++ since it's so much slower 2011-11-02T05:34:28 caution: I guess array.array will have the same time complexity as a list, it's just an array of values instead of pointers to values 2011-11-02T05:34:38 numpy arrays can be allocated really fast 2011-11-02T05:37:26 wow pypy gives my bot a massive speed boost 2011-11-02T05:38:07 add pypy support please 2011-11-02T05:38:18 it's ~5x faster than cpython now I think 2011-11-02T05:38:38 makes my stale tile function run 4x faster 2011-11-02T05:39:10 contestbot, pypy? 2011-11-02T05:39:11 caution: Run as fast as you can and don't look back. 2011-11-02T05:39:54 caution: are you calling the function a lot? 2011-11-02T05:40:18 wow 2011-11-02T05:40:49 i was testing out a 1500 game, took forever each turn 2011-11-02T05:41:20 until turn 750, when something happened, and suddenly my bot just shot through the other 750 turns in 5 seconds without timing out 2011-11-02T05:41:25 confused... 2011-11-02T05:42:02 tcp server? 2011-11-02T05:42:25 nope just local testing with sample bots 2011-11-02T05:42:28 *** antimatroid has quit IRC (Ping timeout: 245 seconds) 2011-11-02T05:42:32 not sure what was taking up all the cpu time 2011-11-02T05:42:38 oh ok 2011-11-02T05:42:47 some of the sample bots are really slow (greedybot) 2011-11-02T05:42:56 OH greedy bot timed out that's why 2011-11-02T05:42:56 yea 2011-11-02T05:42:59 lol 2011-11-02T05:43:04 yeh gonna get rid of that one... takes forever 2011-11-02T05:43:13 at least that means , my bot's explore routine isn't that long! 2011-11-02T05:43:22 make your bot play games against itself :P 2011-11-02T05:43:29 test against your own bots.. old versions to see if they improved 2011-11-02T05:43:41 it goes around in circles still though ==" 2011-11-02T05:48:06 hypothetically speaking, if i had a datastructure that mapped paths between every single walkable tile on a map, how big would that be 2011-11-02T05:48:13 ^^ 2011-11-02T05:50:09 at least I can run the sample bots with pypy to speed up testing 2011-11-02T05:50:13 if you'd make a bfs for every tile covering the whole map and the map would be 200*200(max) then it would be 200^4 :) 2011-11-02T05:50:17 bobbydroptable: Well, how many distinct pairs of tiles are there? 2011-11-02T05:50:43 lets say worst case scenario, a 150x150 map with no water tiles? 2011-11-02T05:50:54 200 is worst case :) 2011-11-02T05:50:59 lol okay 200x200 2011-11-02T05:51:01 That's pretty much (150*150)^2/2, or so. 2011-11-02T05:51:14 Give or take some horrible math, but it's quadratic in size. 2011-11-02T05:51:23 but each tile has a full path to every other tile 2011-11-02T05:51:37 no, you make a distance map for each tile 2011-11-02T05:51:41 so that's a 2d array of ints 2011-11-02T05:51:45 yep 2011-11-02T05:51:55 and it's to large :) 2011-11-02T05:51:56 how would a distance map help 2011-11-02T05:51:59 with a max of 200 2011-11-02T05:52:10 it stores every path to a single target 2011-11-02T05:52:41 every path or every distance? 2011-11-02T05:52:48 o.O? 2011-11-02T05:53:01 6400MB if sizeof (int) is 4 and that's what you're using 2011-11-02T05:53:08 well, you have every distance and thus every path 2011-11-02T05:53:23 hehe.. a sequence of minimal distances is a path :) 2011-11-02T05:53:37 *** Fandekasp has joined #aichallenge 2011-11-02T05:53:40 I think bobbydroptable is getting at you can store every path with 2 bits per tile 2011-11-02T05:53:57 you just have to know which direction to step next 2011-11-02T05:54:04 lol 2011-11-02T05:54:15 something like that 2011-11-02T05:54:20 lots of fun implementing that with a lot of masks 2011-11-02T05:54:27 i'm probably thinking something really really memory inefficient 2011-11-02T05:54:30 and it probably wouldnt be fast =) 2011-11-02T05:54:53 i was thinking path[origin][destination] = ['n','n','n','e','e','e','n','n'] 2011-11-02T05:54:57 lol 2011-11-02T05:55:28 which sounds even worse than any of the other suggestions 2011-11-02T05:55:28 there was a decent post in a pathfinding thread in the forums explaining a decent memory compact way that you could fit all paths precomputed 2011-11-02T05:55:38 its the same as the distance map in this case its even larger because you repeat information 2011-11-02T05:55:42 well gotta run to the bus 2011-11-02T05:55:44 later :) 2011-11-02T05:55:48 lol ok 2011-11-02T05:55:49 janzert: how do they ever find the time to recompute them all? 2011-11-02T05:55:50 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-11-02T05:56:04 when they find new water 2011-11-02T05:56:05 thestinger: incrementally 2011-11-02T05:56:37 not sure anyone is actually doing it though, just someone laid out a method that could possibly work :) 2011-11-02T05:56:50 it seems like you would fall really far behind just recomputing what you need 2011-11-02T05:57:33 if you could implement a way to do it in your spare cycles it might be worth while? 2011-11-02T05:58:01 uh, the official visualizer is broken? 2011-11-02T05:58:11 the top bots are probably usually all their spare cycles on minimax already 2011-11-02T05:58:14 using* 2011-11-02T05:58:35 thestinger: I'd be quite surprised if that is the case 2011-11-02T05:58:48 Minthos: not that I've seen 2011-11-02T05:59:13 does this show correctly then? http://aichallenge.org/visualizer.php?game=47818&user=4823 2011-11-02T05:59:26 seems to be for me 2011-11-02T05:59:29 I just get wall of error text, no game 2011-11-02T05:59:33 *** antimatroid has joined #aichallenge 2011-11-02T05:59:36 :( 2011-11-02T05:59:37 looks ok to me 2011-11-02T05:59:42 try deep refresh 2011-11-02T05:59:46 ctrl F5 2011-11-02T06:02:47 it works, thanks 2011-11-02T06:03:05 *** ikmob has joined #aichallenge 2011-11-02T06:04:31 I present to you the most ridiculous game ending ever: http://ants.fluxid.pl/replay.16827 (notice the "rank stabilized" timing....) 2011-11-02T06:05:27 *** ikmob has quit IRC (Client Quit) 2011-11-02T06:06:29 not so ridiculous 2011-11-02T06:07:24 *** PeterS_ has joined #aichallenge 2011-11-02T06:07:29 orange had ants left on the map so he wasn't eliminated, but he didn't have any hives left so he was screwed 2011-11-02T06:07:36 *** liberforce has joined #aichallenge 2011-11-02T06:07:44 i.e. the rank was stabilized 2011-11-02T06:08:46 So I have no language specific experience with any of the usable languages, guessing C++ is the way to go? 2011-11-02T06:09:11 yes, but the ending score is 3 for me, 0 for him and I should have won, but I didn't 2011-11-02T06:09:15 userjjb: C++ and C is an excellent route to hair loss. 2011-11-02T06:09:21 userjjb: what do you mean the "usable" languages? 2011-11-02T06:09:46 thestinger: the ones that have starter packages 2011-11-02T06:09:53 oh 2011-11-02T06:10:06 Minthos * 2011-11-02T06:10:19 well, you really don't want to start learning C or C++ from scratch for this imo 2011-11-02T06:10:23 I've only programmed in MATLAB, Euphoria, and TI BASIC 2011-11-02T06:11:07 any of the compiled or jvm languages will be more than fast enough - python/ruby are very nice and easy to learn but the performance is hard to deal with 2011-11-02T06:11:39 Venom: I don't know which of them you are, but blue clearly won as he should 2011-11-02T06:12:12 Well I've been meaning to learn C++ so I figured this might be a good impetus to do so 2011-11-02T06:12:12 Minthos: I'm the blue and yea, I got screwed on that one :P 2011-11-02T06:12:27 Venom: no you didn't, you won, what are you complaining about? 2011-11-02T06:12:31 userjjb: well, if you've been planning on learning it go ahead :) 2011-11-02T06:12:34 Is there anything about the task that inherently makes it challenging? 2011-11-02T06:12:45 in C++ specifically 2011-11-02T06:13:03 Minthos: I didn't won. I'm classed #2. 2011-11-02T06:13:12 no, C++ has all the data structures and stuff you could ever want, it's just incredibly verbose/unreadable and it's a huge language so it's hard to learn 2011-11-02T06:13:27 using C would be very, very hard :P 2011-11-02T06:13:50 Venom: it clearly says 1 for me 2011-11-02T06:14:03 Venom: I think your browser is lying to you 2011-11-02T06:14:42 not sure what's going on. I'm confused... nvm 2011-11-02T06:15:19 I noticed under "environment" for C++ in starter packages it has gcc, will I run into issues using Visual Studio? 2011-11-02T06:15:37 userjjb: is you have no reason to specifically learn C++, stay far away from that evil language 2011-11-02T06:15:47 you can use gcc on windows 2011-11-02T06:16:00 *if 2011-11-02T06:16:56 well I already have Visual Studio 2010 Ultimate, so I have all sorts of goodies at my disposal, if I knew how to use them 2011-11-02T06:17:15 userjjb: learn C# instead 2011-11-02T06:17:39 Minthos: why? 2011-11-02T06:17:44 it's a much better language, even I the die-hard linux fan, admits that 2011-11-02T06:17:54 yeah, learn C# or Scala 2011-11-02T06:18:00 hmm 2011-11-02T06:28:34 *** PeterS_ has quit IRC (*.net *.split) 2011-11-02T06:28:34 *** xathis has quit IRC (*.net *.split) 2011-11-02T06:28:34 *** FastEddie_ has quit IRC (*.net *.split) 2011-11-02T06:28:34 *** ^5 <^5!7cb29512@gateway/web/freenode/ip.124.178.149.18> has quit IRC (*.net *.split) 2011-11-02T06:28:34 *** userjjb has quit IRC (*.net *.split) 2011-11-02T06:28:34 *** Venom has quit IRC (*.net *.split) 2011-11-02T06:28:34 *** paulwal has quit IRC (*.net *.split) 2011-11-02T06:28:34 *** asf142 has quit IRC (*.net *.split) 2011-11-02T06:28:34 *** uriahb has quit IRC (*.net *.split) 2011-11-02T06:28:34 *** deroo has quit IRC (*.net *.split) 2011-11-02T06:28:34 *** bobbydroptable has quit IRC (*.net *.split) 2011-11-02T06:28:34 *** ccc has quit IRC (*.net *.split) 2011-11-02T06:28:34 *** capa has quit IRC (*.net *.split) 2011-11-02T06:28:34 *** jnwhiteh has quit IRC (*.net *.split) 2011-11-02T06:28:34 *** spacebat has quit IRC (*.net *.split) 2011-11-02T06:28:34 *** sigh has quit IRC (*.net *.split) 2011-11-02T06:28:35 *** SharkMonkey has quit IRC (*.net *.split) 2011-11-02T06:28:35 *** onensora has quit IRC (*.net *.split) 2011-11-02T06:28:35 *** oddmunds has quit IRC (*.net *.split) 2011-11-02T06:28:35 *** tdubellz has quit IRC (*.net *.split) 2011-11-02T06:28:35 *** Baus has quit IRC (*.net *.split) 2011-11-02T06:28:36 *** kloeri has quit IRC (*.net *.split) 2011-11-02T06:28:36 *** og01 has quit IRC (*.net *.split) 2011-11-02T06:28:37 *** bqf has quit IRC (*.net *.split) 2011-11-02T06:28:37 *** cjoe has quit IRC (*.net *.split) 2011-11-02T06:28:37 *** avdg1 has quit IRC (*.net *.split) 2011-11-02T06:28:38 *** jcdjcd has quit IRC (*.net *.split) 2011-11-02T06:28:38 *** janzert has quit IRC (*.net *.split) 2011-11-02T06:28:38 *** buq2 has quit IRC (*.net *.split) 2011-11-02T06:28:39 *** knyppeldynan has quit IRC (*.net *.split) 2011-11-02T06:28:39 *** Sir_Ragnarok_ has quit IRC (*.net *.split) 2011-11-02T06:28:40 *** v_paul_v_ has quit IRC (*.net *.split) 2011-11-02T06:28:41 *** juuso_ has quit IRC (*.net *.split) 2011-11-02T06:28:42 *** zacharydenton has quit IRC (*.net *.split) 2011-11-02T06:28:42 *** klutometis has quit IRC (*.net *.split) 2011-11-02T06:28:43 *** grom358 has quit IRC (*.net *.split) 2011-11-02T06:28:44 *** rb_ has quit IRC (*.net *.split) 2011-11-02T06:28:44 *** jmcarthur has quit IRC (*.net *.split) 2011-11-02T06:28:44 *** clone1018 has quit IRC (*.net *.split) 2011-11-02T06:28:45 *** Israfel has quit IRC (*.net *.split) 2011-11-02T06:28:46 *** jhawthorn has quit IRC (*.net *.split) 2011-11-02T06:28:46 *** Fluxid has quit IRC (*.net *.split) 2011-11-02T06:28:47 *** jimb0 has quit IRC (*.net *.split) 2011-11-02T06:28:47 *** kire has quit IRC (*.net *.split) 2011-11-02T06:28:48 *** rtward has quit IRC (*.net *.split) 2011-11-02T06:28:48 *** wose has quit IRC (*.net *.split) 2011-11-02T06:28:49 *** liberforce has quit IRC (*.net *.split) 2011-11-02T06:28:49 *** delt0r__ has quit IRC (*.net *.split) 2011-11-02T06:28:49 *** heinrich5991 has quit IRC (*.net *.split) 2011-11-02T06:28:50 *** mawif has quit IRC (*.net *.split) 2011-11-02T06:28:51 *** kronos has quit IRC (*.net *.split) 2011-11-02T06:28:51 *** ejls has quit IRC (*.net *.split) 2011-11-02T06:28:51 *** ruser has quit IRC (*.net *.split) 2011-11-02T06:28:51 *** ybit has quit IRC (*.net *.split) 2011-11-02T06:28:52 *** kingemer has quit IRC (*.net *.split) 2011-11-02T06:28:52 *** Minthos has quit IRC (*.net *.split) 2011-11-02T06:28:53 *** JamesMG has quit IRC (*.net *.split) 2011-11-02T06:28:53 *** jstemmer has quit IRC (*.net *.split) 2011-11-02T06:28:54 *** Lithosphere has quit IRC (*.net *.split) 2011-11-02T06:28:54 *** smoke_carrot has quit IRC (*.net *.split) 2011-11-02T06:28:54 *** isbric has quit IRC (*.net *.split) 2011-11-02T06:28:54 *** aliceinwire has quit IRC (*.net *.split) 2011-11-02T06:28:55 *** AndrewBC has quit IRC (*.net *.split) 2011-11-02T06:28:55 *** krishna has quit IRC (*.net *.split) 2011-11-02T06:28:56 *** timos has quit IRC (*.net *.split) 2011-11-02T06:28:56 *** dscy has quit IRC (*.net *.split) 2011-11-02T06:28:57 *** chris__0076 has quit IRC (*.net *.split) 2011-11-02T06:28:57 *** Alexer has quit IRC (*.net *.split) 2011-11-02T06:28:58 *** Ephemera has quit IRC (*.net *.split) 2011-11-02T06:28:59 *** antonh has quit IRC (*.net *.split) 2011-11-02T06:28:59 *** Zao has quit IRC (*.net *.split) 2011-11-02T06:28:59 *** tobiassjosten has quit IRC (*.net *.split) 2011-11-02T06:28:59 *** f4hy has quit IRC (*.net *.split) 2011-11-02T06:28:59 *** Lerc has quit IRC (*.net *.split) 2011-11-02T06:29:00 *** Acy- has quit IRC (*.net *.split) 2011-11-02T06:29:00 *** wonklab has quit IRC (*.net *.split) 2011-11-02T06:29:01 *** Jak_o_Shadows1 has quit IRC (*.net *.split) 2011-11-02T06:29:02 *** balkemacho has quit IRC (*.net *.split) 2011-11-02T06:29:02 *** Areks_ has quit IRC (*.net *.split) 2011-11-02T06:29:02 *** AxVapor has quit IRC (*.net *.split) 2011-11-02T06:29:02 *** Accoun_ has quit IRC (*.net *.split) 2011-11-02T06:29:03 *** chiller- has quit IRC (*.net *.split) 2011-11-02T06:29:05 *** pgpaskar_ has quit IRC (*.net *.split) 2011-11-02T06:29:06 *** Blkt has quit IRC (*.net *.split) 2011-11-02T06:29:06 *** caution has quit IRC (*.net *.split) 2011-11-02T06:29:06 *** alc has quit IRC (*.net *.split) 2011-11-02T06:29:06 *** Draakon has quit IRC (*.net *.split) 2011-11-02T06:29:07 *** eccentric has quit IRC (*.net *.split) 2011-11-02T06:29:07 *** delt0r has quit IRC (*.net *.split) 2011-11-02T06:29:07 *** jezuz` has quit IRC (*.net *.split) 2011-11-02T06:29:07 *** edcba has quit IRC (*.net *.split) 2011-11-02T06:29:07 *** Darrow has quit IRC (*.net *.split) 2011-11-02T06:29:08 *** a1k0n has quit IRC (*.net *.split) 2011-11-02T06:29:09 *** mayanks43 has quit IRC (*.net *.split) 2011-11-02T06:29:09 *** Garf has quit IRC (*.net *.split) 2011-11-02T06:29:09 *** sig- has quit IRC (*.net *.split) 2011-11-02T06:29:09 *** edgarymr has quit IRC (*.net *.split) 2011-11-02T06:29:10 *** iris1 has quit IRC (*.net *.split) 2011-11-02T06:29:10 *** thestinger has quit IRC (*.net *.split) 2011-11-02T06:29:10 *** Da_Blitz has quit IRC (*.net *.split) 2011-11-02T06:29:10 *** bysin has quit IRC (*.net *.split) 2011-11-02T06:29:11 *** Ttech has quit IRC (*.net *.split) 2011-11-02T06:29:11 *** modafinil has quit IRC (*.net *.split) 2011-11-02T06:29:11 *** moondust has quit IRC (*.net *.split) 2011-11-02T06:29:11 *** rxc has quit IRC (*.net *.split) 2011-11-02T06:29:11 *** krikav has quit IRC (*.net *.split) 2011-11-02T06:29:11 *** welterde has quit IRC (*.net *.split) 2011-11-02T06:29:11 *** acieroid has quit IRC (*.net *.split) 2011-11-02T06:29:11 *** petterw has quit IRC (*.net *.split) 2011-11-02T06:29:11 *** ben---- has quit IRC (*.net *.split) 2011-11-02T06:29:12 *** blueprints has quit IRC (*.net *.split) 2011-11-02T06:29:12 *** Fandekasp has quit IRC (*.net *.split) 2011-11-02T06:29:12 *** Overburn has quit IRC (*.net *.split) 2011-11-02T06:29:13 *** smiley1983 has quit IRC (*.net *.split) 2011-11-02T06:29:14 *** netantho has quit IRC (*.net *.split) 2011-11-02T06:29:14 *** aarossig has quit IRC (*.net *.split) 2011-11-02T06:29:14 *** callahan has quit IRC (*.net *.split) 2011-11-02T06:29:14 *** djstrong has quit IRC (*.net *.split) 2011-11-02T06:29:15 *** ajf|offline has quit IRC (*.net *.split) 2011-11-02T06:29:15 *** dai-ra has quit IRC (*.net *.split) 2011-11-02T06:29:15 *** lericson has quit IRC (*.net *.split) 2011-11-02T06:29:16 *** perryh- has quit IRC (*.net *.split) 2011-11-02T06:29:16 *** nikis has quit IRC (*.net *.split) 2011-11-02T06:29:16 *** j3camero has quit IRC (*.net *.split) 2011-11-02T06:29:16 *** Corey has quit IRC (*.net *.split) 2011-11-02T06:29:17 *** Tumi_ has quit IRC (*.net *.split) 2011-11-02T06:29:17 *** McLeopold has quit IRC (*.net *.split) 2011-11-02T06:29:17 *** AlliedEnvy has quit IRC (*.net *.split) 2011-11-02T06:29:18 *** sorki has quit IRC (*.net *.split) 2011-11-02T06:29:18 *** nplus has quit IRC (*.net *.split) 2011-11-02T06:29:18 *** goffrie has quit IRC (*.net *.split) 2011-11-02T06:29:18 *** rmmh has quit IRC (*.net *.split) 2011-11-02T06:29:18 *** Raimondi has quit IRC (*.net *.split) 2011-11-02T06:29:18 *** Ionic_Groove has quit IRC (*.net *.split) 2011-11-02T06:29:19 *** cbad has quit IRC (*.net *.split) 2011-11-02T06:29:19 *** spm_Draget has quit IRC (*.net *.split) 2011-11-02T06:29:19 *** magiik has quit IRC (*.net *.split) 2011-11-02T06:29:19 *** rofer has quit IRC (*.net *.split) 2011-11-02T06:29:20 *** luizribeiro has quit IRC (*.net *.split) 2011-11-02T06:29:20 *** Zepp has quit IRC (*.net *.split) 2011-11-02T06:29:21 *** jbroman has quit IRC (*.net *.split) 2011-11-02T06:29:21 *** Katiska has quit IRC (*.net *.split) 2011-11-02T06:29:21 *** CIA-24 has quit IRC (*.net *.split) 2011-11-02T06:29:21 *** ChanServ has quit IRC (*.net *.split) 2011-11-02T06:30:20 dropping like flies 2011-11-02T06:32:58 *** g0llum has joined #aichallenge 2011-11-02T06:32:58 *** dorisabayon has joined #aichallenge 2011-11-02T06:32:58 *** spm-Draget has joined #aichallenge 2011-11-02T06:32:58 *** liberforce has joined #aichallenge 2011-11-02T06:32:58 *** Minthos has joined #aichallenge 2011-11-02T06:32:58 *** sigh has joined #aichallenge 2011-11-02T06:32:58 *** Jak_o_Shadows1 has joined #aichallenge 2011-11-02T06:32:58 *** Blkt has joined #aichallenge 2011-11-02T06:32:58 *** cjoe has joined #aichallenge 2011-11-02T06:32:58 *** SharkMonkey has joined #aichallenge 2011-11-02T06:32:59 *** avdg1 has joined #aichallenge 2011-11-02T06:32:59 *** caution has joined #aichallenge 2011-11-02T06:32:59 *** delt0r__ has joined #aichallenge 2011-11-02T06:32:59 *** onensora has joined #aichallenge 2011-11-02T06:32:59 *** balkemacho has joined #aichallenge 2011-11-02T06:32:59 *** grom358 has joined #aichallenge 2011-11-02T06:32:59 *** Areks_ has joined #aichallenge 2011-11-02T06:32:59 *** oddmunds has joined #aichallenge 2011-11-02T06:32:59 *** heinrich5991 has joined #aichallenge 2011-11-02T06:32:59 *** timos has joined #aichallenge 2011-11-02T06:32:59 *** v_paul_v_ has joined #aichallenge 2011-11-02T06:32:59 *** AxVapor has joined #aichallenge 2011-11-02T06:32:59 *** alc has joined #aichallenge 2011-11-02T06:32:59 *** dscy has joined #aichallenge 2011-11-02T06:32:59 *** rb_ has joined #aichallenge 2011-11-02T06:32:59 *** jmcarthur has joined #aichallenge 2011-11-02T06:32:59 *** clone1018 has joined #aichallenge 2011-11-02T06:32:59 *** Accoun_ has joined #aichallenge 2011-11-02T06:32:59 *** chris__0076 has joined #aichallenge 2011-11-02T06:32:59 *** Israfel has joined #aichallenge 2011-11-02T06:32:59 *** tdubellz has joined #aichallenge 2011-11-02T06:32:59 *** mawif has joined #aichallenge 2011-11-02T06:32:59 *** Baus has joined #aichallenge 2011-11-02T06:32:59 *** JamesMG has joined #aichallenge 2011-11-02T06:32:59 *** jcdjcd has joined #aichallenge 2011-11-02T06:32:59 *** Draakon has joined #aichallenge 2011-11-02T06:32:59 *** og01 has joined #aichallenge 2011-11-02T06:32:59 *** Alexer has joined #aichallenge 2011-11-02T06:32:59 *** eccentric has joined #aichallenge 2011-11-02T06:32:59 *** janzert has joined #aichallenge 2011-11-02T06:32:59 *** buq2 has joined #aichallenge 2011-11-02T06:32:59 *** bqf has joined #aichallenge 2011-11-02T06:32:59 *** kloeri has joined #aichallenge 2011-11-02T06:32:59 *** Sir_Ragnarok_ has joined #aichallenge 2011-11-02T06:32:59 *** jstemmer has joined #aichallenge 2011-11-02T06:32:59 *** delt0r has joined #aichallenge 2011-11-02T06:32:59 *** Lithosphere has joined #aichallenge 2011-11-02T06:32:59 *** knyppeldynan has joined #aichallenge 2011-11-02T06:32:59 *** smoke_carrot has joined #aichallenge 2011-11-02T06:32:59 *** Zao has joined #aichallenge 2011-11-02T06:32:59 *** zacharydenton has joined #aichallenge 2011-11-02T06:32:59 *** juuso_ has joined #aichallenge 2011-11-02T06:32:59 *** Ephemera has joined #aichallenge 2011-11-02T06:32:59 *** isbric has joined #aichallenge 2011-11-02T06:32:59 *** ybit has joined #aichallenge 2011-11-02T06:32:59 *** jezuz` has joined #aichallenge 2011-11-02T06:32:59 *** chiller- has joined #aichallenge 2011-11-02T06:32:59 *** edcba has joined #aichallenge 2011-11-02T06:32:59 *** antonh has joined #aichallenge 2011-11-02T06:32:59 *** Darrow has joined #aichallenge 2011-11-02T06:32:59 *** aliceinwire has joined #aichallenge 2011-11-02T06:32:59 *** jhawthorn has joined #aichallenge 2011-11-02T06:32:59 *** jimb0 has joined #aichallenge 2011-11-02T06:32:59 *** tobiassjosten has joined #aichallenge 2011-11-02T06:32:59 *** f4hy has joined #aichallenge 2011-11-02T06:32:59 *** kronos has joined #aichallenge 2011-11-02T06:32:59 *** wose has joined #aichallenge 2011-11-02T06:32:59 *** ejls has joined #aichallenge 2011-11-02T06:32:59 *** krishna has joined #aichallenge 2011-11-02T06:32:59 *** Fluxid has joined #aichallenge 2011-11-02T06:32:59 *** klutometis has joined #aichallenge 2011-11-02T06:32:59 *** mayanks43 has joined #aichallenge 2011-11-02T06:32:59 *** kingemer has joined #aichallenge 2011-11-02T06:32:59 *** Lerc has joined #aichallenge 2011-11-02T06:32:59 *** Acy- has joined #aichallenge 2011-11-02T06:32:59 *** ruser has joined #aichallenge 2011-11-02T06:32:59 *** AndrewBC has joined #aichallenge 2011-11-02T06:32:59 *** rtward has joined #aichallenge 2011-11-02T06:32:59 *** spacebat has joined #aichallenge 2011-11-02T06:32:59 *** jnwhiteh has joined #aichallenge 2011-11-02T06:32:59 *** wonklab has joined #aichallenge 2011-11-02T06:32:59 *** a1k0n has joined #aichallenge 2011-11-02T06:32:59 *** pgpaskar_ has joined #aichallenge 2011-11-02T06:32:59 *** kire has joined #aichallenge 2011-11-02T06:33:06 *** edgarymr has joined #aichallenge 2011-11-02T06:33:06 *** iris1 has joined #aichallenge 2011-11-02T06:33:06 *** thestinger has joined #aichallenge 2011-11-02T06:33:06 *** Da_Blitz has joined #aichallenge 2011-11-02T06:33:06 *** bysin has joined #aichallenge 2011-11-02T06:33:06 *** Ttech has joined #aichallenge 2011-11-02T06:33:06 *** modafinil has joined #aichallenge 2011-11-02T06:33:06 *** moondust has joined #aichallenge 2011-11-02T06:33:06 *** rxc has joined #aichallenge 2011-11-02T06:33:06 *** krikav has joined #aichallenge 2011-11-02T06:33:06 *** welterde has joined #aichallenge 2011-11-02T06:33:06 *** acieroid has joined #aichallenge 2011-11-02T06:33:06 *** blueprints has joined #aichallenge 2011-11-02T06:33:06 *** ben---- has joined #aichallenge 2011-11-02T06:33:25 *** lericson has joined #aichallenge 2011-11-02T06:33:30 *** j3camero has joined #aichallenge 2011-11-02T06:33:42 *** nikis has joined #aichallenge 2011-11-02T06:33:45 *** Tumi_ has joined #aichallenge 2011-11-02T06:33:47 *** magiik has joined #aichallenge 2011-11-02T06:33:50 *** sorki has joined #aichallenge 2011-11-02T06:33:55 *** rmmh has joined #aichallenge 2011-11-02T06:33:57 *** perryh- has joined #aichallenge 2011-11-02T06:34:00 *** ajf|offline has joined #aichallenge 2011-11-02T06:34:01 *** Katiska has joined #aichallenge 2011-11-02T06:34:13 *** Zepp has joined #aichallenge 2011-11-02T06:34:16 *** netantho has joined #aichallenge 2011-11-02T06:34:17 *** cbad has joined #aichallenge 2011-11-02T06:34:18 *** djstrong has joined #aichallenge 2011-11-02T06:34:24 *** aarossig has joined #aichallenge 2011-11-02T06:35:43 *** AlliedEnvy has joined #aichallenge 2011-11-02T06:36:32 *** nplus has joined #aichallenge 2011-11-02T06:37:01 *** luizribeiro has joined #aichallenge 2011-11-02T06:37:57 *** Overburn has joined #aichallenge 2011-11-02T06:38:24 *** FastEddie_ has joined #aichallenge 2011-11-02T06:39:13 *** Jak_o_Shadows has joined #aichallenge 2011-11-02T06:39:48 *** Jak_o_Shadows1 has quit IRC (Ping timeout: 240 seconds) 2011-11-02T06:43:13 *** edgarymr has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 7.0.1/20110928134238]) 2011-11-02T06:45:06 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-11-02T06:47:10 *** CIA-53 has joined #aichallenge 2011-11-02T06:49:06 *** smiley1983 has joined #aichallenge 2011-11-02T06:50:43 has anyone tried their hand at a self learning algorithm? 2011-11-02T06:51:19 would it be illegal to let your bot connect through tcp to a separate server to export game state and import instructions? 2011-11-02T06:51:41 *** bobbydroptable has joined #aichallenge 2011-11-02T06:51:41 *** McLeopold has joined #aichallenge 2011-11-02T06:51:41 *** ChanServ has joined #aichallenge 2011-11-02T06:51:41 *** card.freenode.net sets mode: +o ChanServ 2011-11-02T06:53:14 *** bobbydroptable has quit IRC (Client Quit) 2011-11-02T06:53:23 caution: start from scratch 2011-11-02T06:53:42 I am 2011-11-02T06:53:48 :) 2011-11-02T06:55:07 does the game server have internet access eddie? 2011-11-02T06:55:44 that definitely isn't within the rules 2011-11-02T06:55:54 no, you cannot use network 2011-11-02T06:56:06 you could learn by automatically processing the replays from the website instead 2011-11-02T06:56:17 use the tcp server for stuff like that 2011-11-02T06:56:21 and you cannot store state between battles in any way 2011-11-02T06:57:02 No network, no files, etc. 2011-11-02T06:57:42 If you're going to do some kind of genetic or learning bot, teach it locally or via replays. 2011-11-02T06:58:00 *** foRei has joined #aichallenge 2011-11-02T06:58:09 ok thanks.. i thought it would be illegal but i couldnt find it in the rules 2011-11-02T06:58:40 *** nikis has quit IRC (Changing host) 2011-11-02T06:58:40 *** nikis has joined #aichallenge 2011-11-02T06:58:40 *** djstrong has quit IRC (Changing host) 2011-11-02T06:58:40 *** djstrong has joined #aichallenge 2011-11-02T06:59:12 well, writing to a socket is writing to a file :) 2011-11-02T06:59:45 *** sig- has joined #aichallenge 2011-11-02T07:00:03 true :) 2011-11-02T07:00:19 *** bobbydroptable has joined #aichallenge 2011-11-02T07:00:37 freenode webchat totally just screwed up my web browser ==" 2011-11-02T07:01:15 guess i'll just have to learn this irc chat client thingie 2011-11-02T07:02:03 *** Migi32 has joined #aichallenge 2011-11-02T07:02:10 it would be interesting to see this competition but without letting ants share information with eachother 2011-11-02T07:02:41 FastEddie_ how would they not be able to share information with one another? 2011-11-02T07:02:41 *** Garf has joined #aichallenge 2011-11-02T07:03:05 *** ajf|offline is now known as Guest53254 2011-11-02T07:03:05 *** luizribeiro is now known as Guest628 2011-11-02T07:03:11 well if you let the bot control only one ant at the time, and then call multiple instances of the bot for each ant that you own 2011-11-02T07:03:27 you could only share information in the game world for example by doing a bee like dance 2011-11-02T07:03:44 or maybe even allow ants to drop different scents on the map 2011-11-02T07:03:56 ooh that would be a completey different dynamic yeh 2011-11-02T07:04:01 spawn a process when an ant spawns, kill a process when an ant dies, I like it 2011-11-02T07:04:01 right now all the ants know everything because there is one overarching bot that is in full control 2011-11-02T07:05:21 *** jbroman has joined #aichallenge 2011-11-02T07:06:12 i'm not suggeting that this gets changed of course but i just thought it was an interesting idea that would come up with more lifelike emergent swarming 2011-11-02T07:06:25 would definitely be interesting =) 2011-11-02T07:09:08 *** alc has quit IRC (Ping timeout: 245 seconds) 2011-11-02T07:09:49 *** pedrosorio has joined #aichallenge 2011-11-02T07:10:08 I like it as well! 2011-11-02T07:10:13 *** cyphase has joined #aichallenge 2011-11-02T07:10:29 When I first read about the challenge I thought it would be similar to that. 2011-11-02T07:17:07 *** Raimondi has joined #aichallenge 2011-11-02T07:18:30 I'm not too sure about that. You could let it go to food, explore randomly and let it do relatively simple and uncoordinated battles. 2011-11-02T07:18:46 no time for cute bee dances when you're in the middle of combat ;) 2011-11-02T07:19:24 so all in all I think the resulting bots would be less interesting 2011-11-02T07:19:59 but maybe it's worth a try for the next contest or something 2011-11-02T07:21:02 *** marijnfs has joined #aichallenge 2011-11-02T07:21:08 You have to introduce additional concepts, like pheromones (e.g. creating trails other ants can see). 2011-11-02T07:21:55 hello, does someone know how to check why a bot crashes in a game on the server? 2011-11-02T07:22:16 my ants stop moving after several turn, which i guess means they ran out their time limit 2011-11-02T07:22:21 does it crash when you run it locally? 2011-11-02T07:22:24 but there is no further feedback 2011-11-02T07:22:28 no only on the site 2011-11-02T07:22:41 and locally it seems to be fast enough 2011-11-02T07:22:44 but my pc is quite fast 2011-11-02T07:22:56 my ai runs much slower on the server 2011-11-02T07:24:14 *** userjjb has joined #aichallenge 2011-11-02T07:24:35 *** mleise has joined #aichallenge 2011-11-02T07:25:20 *** pedrosorio has quit IRC (Ping timeout: 265 seconds) 2011-11-02T07:25:22 marijnfs is it a crash or a timeout, do you know? 2011-11-02T07:25:40 so i dont know, there is no feedback 2011-11-02T07:25:45 or i can't find it\ 2011-11-02T07:26:10 in the replay it should say crash or timeout in the bar above 2011-11-02T07:26:11 i am guessing a timeout but i dont know how close i am to it 2011-11-02T07:26:15 ok 2011-11-02T07:26:20 lets see 2011-11-02T07:26:30 and you can test it locally by setting the turn limit much lower 2011-11-02T07:27:13 ah ok it says timeout 2011-11-02T07:27:23 but locally it seems fast enough 2011-11-02T07:27:47 i guess i need to be more efficient 2011-11-02T07:28:00 the server's turntime is 500, but by default in a lot of the test scripts its 1000. 2011-11-02T07:28:08 so first I would make sure your turntime is 500 2011-11-02T07:28:31 euhm .25 seconds right? 2011-11-02T07:28:35 500 what? 2011-11-02T07:28:45 milliseconds - so 0.5 seconds 2011-11-02T07:29:33 i recommend checking time remaining to make sure you have a buffer, say 10ms, before doing anything complex (like a path find). 2011-11-02T07:29:39 If there's not enough time left, just skip it. 2011-11-02T07:30:42 In practice, I found edge cases would randomly cause timeouts. Been better since I started actually working with the clock. :) 2011-11-02T07:30:44 locally it doesnt even timeout with time limit=.01 seconds 2011-11-02T07:30:58 using a clock is a good idea;) 2011-11-02T07:31:06 what map are you using locally? 2011-11-02T07:31:13 use one of the big mazes 2011-11-02T07:31:28 *** paulwal has joined #aichallenge 2011-11-02T07:31:37 *** olexs has joined #aichallenge 2011-11-02T07:32:05 *** adhlssu07 has joined #aichallenge 2011-11-02T07:32:10 whats a bigger one 2011-11-02T07:32:16 i use the map that caused the timeout online 2011-11-02T07:33:12 i also write to a file for debugging 2011-11-02T07:33:21 i dont know how that is handled on the server 2011-11-02T07:34:40 you are not allowed to write to files afaik :) 2011-11-02T07:34:51 perhaps a stupid question, but are you catching uncaught exceptions? 2011-11-02T07:35:04 if so a crash would likely produce a "timeout" as far as the server knows. 2011-11-02T07:36:46 i use go 2011-11-02T07:36:48 The current phase of the contest will end December 18th 2011-11-02T07:36:49 no exceptions 2011-11-02T07:37:23 i reuploaded my code 2011-11-02T07:37:29 to see if i maybe send the wrong one 2011-11-02T07:37:40 if it still timeouts i'll try removing the file stuff 2011-11-02T07:37:48 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-11-02T07:37:51 too bad you have to wait so long for games 2011-11-02T07:38:18 168.0 minutes :O 2011-11-02T07:38:27 yeah. it's the one frustration in an otherwise extremely well done contest imo. 2011-11-02T07:39:09 it's way faster now that they added deactivation, before I would have to wait 1000 minutes for a game :\ 2011-11-02T07:39:17 hehe 2011-11-02T07:39:26 yeah they added some stopping rules 2011-11-02T07:39:35 *** ama2er has joined #aichallenge 2011-11-02T07:39:48 maybe on the lower ranks they should start with small battles on small maps 2011-11-02T07:40:13 and some standard tests to see if you can collect food in a small map 2011-11-02T07:41:45 what is your current strategy 2011-11-02T07:42:02 Three days is a long time to let a start package run and waste cycles. I think if your bot fails at say collecting food X games in a row it gets automatically de-activated. 2011-11-02T07:42:28 *it should get automatically de-activated, etc. 2011-11-02T07:42:34 But definitely glad to see them improving it. 2011-11-02T07:42:51 what are the memory constraints for a bot? 2011-11-02T07:43:03 i've heard it to be 1Gb 2011-11-02T07:43:33 *** Garf has quit IRC (Quit: Make a new plan, Stan!) 2011-11-02T07:44:11 *** ama2er has quit IRC (Ping timeout: 265 seconds) 2011-11-02T07:45:10 i guess to attract people, they like you to be able to send your starter package and see ants move 2011-11-02T07:45:26 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Ping timeout: 258 seconds) 2011-11-02T07:46:45 *** Garf has joined #aichallenge 2011-11-02T07:48:33 Yeah it's really smart. The instant feedback/progress is encouraging. It just shouldn't be at the detriment to other players. Like, if a person uploads an untouched starter bot (based on a hash) no need to run tha thing often. :) 2011-11-02T07:49:00 detect the starter package and show a premade replay lol 2011-11-02T07:49:50 is it definitely the starter packages that make it so long between games? 2011-11-02T07:51:47 well, the game- terminate conditions go some way towards shortening noobie games 2011-11-02T07:51:57 like "food not being gathered" 2011-11-02T07:52:15 isn't the amount of games you play related to your rating? 2011-11-02T07:52:46 I believe so. But first time uploads appear to have their own queue, so not entirely. 2011-11-02T07:54:27 I guess the only thing that really frustrates me is the penalty when uploading a new version of a bot. That cycle of upload a new version, early skill rating leads to very few games ... finally play some real games, see things that need to be improved. Repeat. 2011-11-02T07:55:07 just use the tcp server for testing then 2011-11-02T07:55:20 *** Jak_o_Shadows has quit IRC (Remote host closed the connection) 2011-11-02T07:55:42 @tcp 2011-11-02T07:55:43 thestinger: tcp could be http://ants.fluxid.pl/howto. 2011-11-02T07:56:11 yeah, but playing only 1 game per 6 hours is quite discouraging. I wonder if xathis got his skill reset to 0, how many days it would take to get back to the top 2011-11-02T07:56:22 *** rofer has joined #aichallenge 2011-11-02T07:57:04 *** amstan_ has joined #aichallenge 2011-11-02T07:57:04 *** ChanServ sets mode: +o amstan_ 2011-11-02T07:57:05 I'm noobish, how does the TCP server help / what is it? 2011-11-02T07:57:19 you can use it play other real people 2011-11-02T07:57:22 +to 2011-11-02T07:57:40 without a big wait time between games like the aichallenge server 2011-11-02T07:57:47 balkemacho, you run your bot locally, it talks to a server and sends its orders over the internet 2011-11-02T07:58:09 that way you get one game per 5 minutes instead of 5 hours 2011-11-02T07:58:14 Oh. Damn. I did not know about that. :) 2011-11-02T07:58:22 What are the username password args? Do you have to register somewhere? 2011-11-02T07:58:41 the aichallenge account credentials i presume 2011-11-02T07:59:09 I wouldn't think you'd want to give your aichallenge login details to a third party server 2011-11-02T07:59:18 no, if the username isn't taken yet, it registers that username/password combination for you 2011-11-02T07:59:24 cool 2011-11-02T07:59:33 oh i see. 2011-11-02T07:59:43 i didnt realize this was a third party server. i get it now. 2011-11-02T08:01:12 *** og01 has quit IRC (Ping timeout: 256 seconds) 2011-11-02T08:01:16 *** og01 has joined #aichallenge 2011-11-02T08:02:40 nice ending: http://ants.fluxid.pl/replay.16918 2011-11-02T08:03:44 *** Accoun_ is now known as Accoun 2011-11-02T08:04:49 *** UncleVasya has joined #aichallenge 2011-11-02T08:06:20 haha unkillable hill 2011-11-02T08:08:00 three in width should have done it, right? 2011-11-02T08:08:48 lol strcat playing against himself 2011-11-02T08:08:54 and one times out 2011-11-02T08:08:58 *** amstan_ is now known as amstan 2011-11-02T08:09:48 would three in width do it or not? 2011-11-02T08:10:07 hey how does one private message another person? 2011-11-02T08:10:41 try /msg nickname message 2011-11-02T08:10:50 or /query nickname 2011-11-02T08:12:24 cool works 2011-11-02T08:13:59 *** Palmik has joined #aichallenge 2011-11-02T08:16:58 bobbydroptable: yeah, I accidentally had them writing to the same log file and it totally screwed up 2011-11-02T08:17:20 you also write to logfiles on the server? 2011-11-02T08:17:36 i am wondering if that is what is timing out my bot 2011-11-02T08:17:43 no, locally 2011-11-02T08:17:53 you cannot write to files on the server, it's against the rules 2011-11-02T08:18:38 hmm 2011-11-02T08:19:01 i wonder what my bot is doing then 2011-11-02T08:19:16 maybe the writes just fail on an unopened file 2011-11-02T08:20:14 *** bob_ has joined #aichallenge 2011-11-02T08:20:19 *** bob_ is now known as othernickthen 2011-11-02T08:20:36 Is anyone working on getting the c package to support hills? 2011-11-02T08:20:38 mmm so my bot explores the map now 2011-11-02T08:20:46 but all the other ants follow the same damn paths lol 2011-11-02T08:20:50 instead of splitting up 2011-11-02T08:20:54 *** PeterS_ has joined #aichallenge 2011-11-02T08:21:01 hehe 2011-11-02T08:21:04 Is it ok, to upload a bot that prints a bunch of crap to stderr? 2011-11-02T08:21:14 like a parade 2011-11-02T08:21:16 yeah but it will probably time out if you spam stuff to stderr 2011-11-02T08:21:39 I thought stderr was counted as compilation errors? 2011-11-02T08:22:24 you might be right 2011-11-02T08:22:45 stderr as compilation errors? 2011-11-02T08:22:47 I'll just clean it up 2011-11-02T08:23:07 my bot writes to stderr and works just fine 2011-11-02T08:23:14 meh... I can't tell if the new or old version of my bot is better 2011-11-02T08:23:20 yeah i cant imagine that being a problem 2011-11-02T08:23:30 I'm trying to figure out foraging (revisiting explored locations after they haven't been seen for a while) 2011-11-02T08:23:50 play them against each other 2011-11-02T08:23:54 I am :) 2011-11-02T08:24:19 new: http://ants.fluxid.pl/player/strcat_next old: http://ants.fluxid.pl/player/strcat 2011-11-02T08:25:38 *** v_paul_v_ has quit IRC (Remote host closed the connection) 2011-11-02T08:25:51 *** v_paul_v_ has joined #aichallenge 2011-11-02T08:26:43 does that site do ranking and stuff? 2011-11-02T08:27:21 yeah, but it's meaningless because there aren't enough bots 2011-11-02T08:27:52 my bot jumps between rank 30 and rank 150 from game to game 2011-11-02T08:28:40 hmm maybe the rankings should have a little more stable logic 2011-11-02T08:28:56 the individual games are a lot more useful than the rankings on the tcp server imo 2011-11-02T08:28:58 who is the server courtesy of? 2011-11-02T08:29:09 Fluxid is hosting that one I guess 2011-11-02T08:29:17 *** formallyknownasd has joined #aichallenge 2011-11-02T08:29:21 someone else made the tcp server/client 2011-11-02T08:29:29 yup 2011-11-02T08:30:19 thestinger: http://ants.fluxid.pl/replay.16918 awesome! :D 2011-11-02T08:30:46 lol :) 2011-11-02T08:31:52 I gave my ants some fear of the enemy but it's not enough yet so they still suicide 2011-11-02T08:32:13 I might have an off-by-one error 2011-11-02T08:32:54 i think that if one made 5x1 row of ants at 108,111 on this map 2011-11-02T08:32:59 this would be unbeatable 2011-11-02T08:33:09 Shouldn't be a row 2011-11-02T08:33:13 should curve 2011-11-02T08:33:19 corners are bad 2011-11-02T08:33:37 othernickthen: look at 108,111 there 2011-11-02T08:33:46 no way to break through wall 2011-11-02T08:33:55 because wall corners are in water 2011-11-02T08:34:41 do i have to look for my name in the rankings? 2011-11-02T08:34:47 or is there a page that lists only my games? 2011-11-02T08:34:51 on fluxid 2011-11-02T08:35:06 marijnfs: just klick on your name 2011-11-02T08:35:19 if could come up with algorithm which could find such places... 2011-11-02T08:36:31 *** othernickthen has quit IRC (Quit: Page closed) 2011-11-02T08:36:53 *** bumhole has joined #aichallenge 2011-11-02T08:37:09 actually, curve or line don't matter 2011-11-02T08:37:11 its sounds so easy 2011-11-02T08:37:23 i've done some sample scenarios to try out different defenses 2011-11-02T08:37:54 on those scaled it becomes more like chess 2011-11-02T08:38:03 you could do some local mini-max perhaps 2011-11-02T08:38:15 s 2011-11-02T08:38:49 I think my ants are just determined to kill themselves sometimes... 2011-11-02T08:39:29 *** ping-- has joined #aichallenge 2011-11-02T08:40:07 *** Ionic_Groove has joined #aichallenge 2011-11-02T08:40:07 how is the "Next game should be within x minutes" calculated? 2011-11-02T08:41:11 how is the mu/sigma calculated on your server Fluxid? 2011-11-02T08:41:14 *** grom358 has quit IRC (Quit: Leaving) 2011-11-02T08:41:28 last n games? 2011-11-02T08:41:32 arcs actually increase the number of corners exposed 2011-11-02T08:41:39 i would rank based on the mu 2011-11-02T08:41:42 more stable 2011-11-02T08:42:05 if you have a line with depth, you will find that the inner ants die, and it turns into an arc anyway 2011-11-02T08:43:24 *** liberforce has quit IRC (Quit: Leaving.) 2011-11-02T08:45:07 *** yoden has joined #aichallenge 2011-11-02T08:46:10 *** Rinum has joined #aichallenge 2011-11-02T08:46:35 I was wondering... how can we know (though code) if a hill has been razed? 2011-11-02T08:47:50 Rinum: I do following: 2011-11-02T08:48:10 *** liberforce has joined #aichallenge 2011-11-02T08:48:49 I keep list of hills between turn. In the begining of the turn I check every hill in list: 2011-11-02T08:48:51 *** ikaros has joined #aichallenge 2011-11-02T08:49:14 if a known hill is in vision but is no longer a hill 2011-11-02T08:49:30 ahhh that works 2011-11-02T08:49:41 does it just become land? 2011-11-02T08:50:05 Rinum: it can be land or Ant 2011-11-02T08:50:12 *** dai-ra has joined #aichallenge 2011-11-02T08:50:17 where Ant owner != hill owner 2011-11-02T08:50:20 ok, thanks 2011-11-02T08:51:12 from what I can tell in the python starter pack 2011-11-02T08:51:25 it clears the hills data at every turn 2011-11-02T08:51:43 meaning that if it is in vision, and the hill isn't there, that means it's gone. 2011-11-02T08:51:47 And so in every starterpack 2011-11-02T08:52:17 yeh just wanted to make it clear. because a hill in vision still gives you LAND. 2011-11-02T08:52:24 *** callahan has joined #aichallenge 2011-11-02T08:56:20 *** BoAnd has joined #aichallenge 2011-11-02T08:56:59 *** conor_f has joined #aichallenge 2011-11-02T08:57:03 What does '#37' means in the Opponents column at the Laest Games table? 2011-11-02T08:57:14 *Last Games 2011-11-02T08:57:30 *Latest Games 2011-11-02T08:57:41 does the fluxid script only run one game and then exit? 2011-11-02T08:57:51 how can i make it run continually 2011-11-02T08:58:08 *** paulwal has quit IRC (Ping timeout: 265 seconds) 2011-11-02T08:58:21 -1 as the last parametr 2011-11-02T08:58:30 *** ^5 <^5!7cb29512@gateway/web/freenode/ip.124.178.149.18> has joined #aichallenge 2011-11-02T08:58:52 <^5> Woo, my bot actually works half decently 2011-11-02T09:00:41 :) 2011-11-02T09:01:07 well done 2011-11-02T09:01:11 *** patrisk has joined #aichallenge 2011-11-02T09:01:47 amstan: hi. How about add Deactivated Submissions Count to the server_stats page? And, if possible, total time saved due to deactivating. 2011-11-02T09:02:04 UncleVasya: add an issue? 2011-11-02T09:02:54 no, thanks :) 2011-11-02T09:03:32 how many people here are doing the stanford intro to ai? 2011-11-02T09:03:41 conor_f: me 2011-11-02T09:03:56 what do you make of it? I think it's very poorely done tbh :/ 2011-11-02T09:04:12 conor_f: yeah, it got worse 2011-11-02T09:04:17 conor_f: i really liked the first chapter 2011-11-02T09:04:22 and maybe even chapter 3 2011-11-02T09:04:28 I had good expectations for it, but they're gone now :( 2011-11-02T09:04:31 but then it's just this whole mess that's not explained 2011-11-02T09:05:04 yeah, exactly. It's fine if you did maths in uni/college/whatever, but if you had never dealt with probability before, it was awful :P 2011-11-02T09:05:29 he just skips over really important parts :/ 2011-11-02T09:05:36 *** aerique has joined #aichallenge 2011-11-02T09:05:43 hi guys 2011-11-02T09:05:48 hey 2011-11-02T09:05:54 the machine learning class is very nice imho 2011-11-02T09:06:02 Garf: i heard that from reddit 2011-11-02T09:06:04 could someone tell me how people debug their bot ? 2011-11-02T09:06:09 You can learn the whole math course right in this channel. Talk to antimatroid :D 2011-11-02T09:06:16 attaching gdb to a C ot doesn't seem to work 2011-11-02T09:06:37 Garf: same, what are the programming problems like for it? 2011-11-02T09:06:40 *to a C bot 2011-11-02T09:06:57 Garf: was going to do it, but would have been too much for me to catch up on 2011-11-02T09:06:59 programming various classifier thingies in Matlab/Octave 2011-11-02T09:07:11 mostly matrix math so far 2011-11-02T09:07:21 ooh, thank god I didn't do it so :) 2011-11-02T09:07:30 *** xathis has joined #aichallenge 2011-11-02T09:07:48 hum... how old are people here ? :p 2011-11-02T09:07:55 19 2011-11-02T09:07:56 *** adhlssu07 has quit IRC (Quit: adhlssu07) 2011-11-02T09:07:58 32 2011-11-02T09:08:05 others ? 2011-11-02T09:08:05 I had no problem understanding anything, the maths is fairly light and very functional 2011-11-02T09:08:06 liberforce: I've just started this challenge, but isn't there a section on debugging on the site/readme? 2011-11-02T09:08:12 liberforce: mostly univ/college level, but others are welcome too 2011-11-02T09:08:18 16 :P 2011-11-02T09:08:26 i'm 19 2011-11-02T09:08:27 i'm pretty sure McLeopold and janzert are much older 2011-11-02T09:08:48 *** olexs has quit IRC (Quit: Leaving.) 2011-11-02T09:08:56 conor_f: i've consulted the development tools section of the forum, but found nothing for my problem 2011-11-02T09:09:23 liberforce: depends on the language 2011-11-02T09:09:23 20 2011-11-02T09:09:40 cat README.txt | grep debug 2011-11-02T09:10:02 conor_f: i don't think so 2011-11-02T09:10:15 amstan: i'm running a C bot 2011-11-02T09:10:25 *** olexs has joined #aichallenge 2011-11-02T09:10:27 shows Bug.h which I presume is used for debugging 2011-11-02T09:10:32 liberforce: i believe you can attach gdb to already running processes 2011-11-02T09:10:33 liberforce: I'm using C++ 2011-11-02T09:11:06 conor_f: that's just for printing things, it's in a header file so you can easily take it out for submitting 2011-11-02T09:11:12 liberforce: amstan: depends on what sort of debuggins is being done I suppose 2011-11-02T09:11:28 *debugging 2011-11-02T09:11:43 well, i attach it after calling playgame.py, but gdb can't print the backtrace nor list the source 2011-11-02T09:11:46 liberforce: if you want to the attaching business, you have to change your timeout params so it doesn't timeout when gdb reaches a breakpoint 2011-11-02T09:12:33 liberforce: that's the extent of the stuff i know. in most debugging i do, the process is started by gdb 2011-11-02T09:12:36 *** argiopeweb has joined #aichallenge 2011-11-02T09:13:03 amstan: i'd prefer to directly run gdb and tell playgame to reuse my process... 2011-11-02T09:13:10 do you know how to do that ? 2011-11-02T09:13:14 Huh 2011-11-02T09:13:20 liberforce: i don't think you can 2011-11-02T09:13:26 Any chance of y'all installing Data.Vector for those of us programming in Haskell? 2011-11-02T09:13:46 amstan: so how do you debug by directly running gdb ? 2011-11-02T09:13:46 can someone tell me why 'strcat' wins in this game: http://ants.fluxid.pl/replay.16992 2011-11-02T09:13:51 argiopeweb: probably, but you might have to do it yourself 2011-11-02T09:14:01 liberforce: gdb execname 2011-11-02T09:14:07 ow wait 2011-11-02T09:14:14 watching wrong color 2011-11-02T09:14:16 liberforce: but that entails that gdb steals your stdin, and that won't work for playgame.py 2011-11-02T09:14:20 great stuff if your colorblind;) 2011-11-02T09:14:30 marijnfs: yeah, we tried to keep that in mind 2011-11-02T09:14:46 amstan: Do it myself? On your servers? 2011-11-02T09:14:59 @repo 2011-11-02T09:15:00 amstan: repo could be https://github.com/aichallenge/aichallenge. 2011-11-02T09:15:11 argiopeweb: see setup/worker_setup.py 2011-11-02T09:15:12 Ah, it all becomes clear. 2011-11-02T09:15:30 amstan: hum... if gdb steals your stdin, how do you debug with ? 2011-11-02T09:15:37 is a dictionary in PY the same as a hash/associative array in Perl/PHP? 2011-11-02T09:15:51 nice, I think I actually made a big improvement to my bot (strcat_next is destroying strcat now :D) 2011-11-02T09:15:56 liberforce: most of the stuff i have doesn't care about stdin very much 2011-11-02T09:16:12 amstan: is the same problem of stdin stealling occurring when attaching gdb to the process ? 2011-11-02T09:16:32 amstan: but parsing the server input requires stdin... 2011-11-02T09:16:54 by stdin i mean that gdb starts and it does it's thing, shows you a prompt and all that 2011-11-02T09:17:04 you don't want playgame to deal with that 2011-11-02T09:17:15 liberforce: as i said, what you're looking for is to attach gdb to the already running process(started by playgame) 2011-11-02T09:17:42 that way gdb will have its own stdin/stdout(connected to your terminal), and your bot's stdin/stdout will go to playgame 2011-11-02T09:17:43 amstan: ok, so what I need to do is augment the timeout 2011-11-02T09:17:49 playgame.py --help 2011-11-02T09:18:16 there's 2 timeouts, 1000ms(turn time) and 3000ms(start) 2011-11-02T09:18:23 forgot how they're called 2011-11-02T09:18:31 amstan: i read the help, but the symptoms I had on gdb were quite strange and I didn't thought of timeout right away 2011-11-02T09:18:50 *** SharkMonkey has quit IRC (Ping timeout: 256 seconds) 2011-11-02T09:19:21 amstan: loadtime and turntime 2011-11-02T09:19:28 amstan: Not sure I can help you from there. Doesn't look like y'all are installing anything other than the base Haskell Platform stuff. 2011-11-02T09:22:01 *** Areks_ has left #aichallenge ("Once you know what it is you want to be true, instinct is a very useful device for enabling you to know that it is") 2011-11-02T09:22:08 conor_f: a dictionary is a hash table, there are also ordered dictionaries in the standard library 2011-11-02T09:22:12 those are some kind of tree I guess 2011-11-02T09:22:38 so a cpp equivalent? Was going to use some form of vector but can't now 2011-11-02T09:22:47 *** bumhole has quit IRC (Ping timeout: 265 seconds) 2011-11-02T09:23:06 I'm just following along with the tutorials and I see a dictionary data structure as all the code is done in PY 2011-11-02T09:23:20 c++ has ordered_map and ordered_set (which are like dict() and set()) if you use c++11 (which you can, just use MyBot.cpp instead of .cc) 2011-11-02T09:23:21 on the canvas, are the ants dead on the turn the lines join the two ants or the turn after? 2011-11-02T09:23:29 *** edcba has left #aichallenge 2011-11-02T09:23:45 unordered_map and unordered_set* lol 2011-11-02T09:24:10 c++ set and map are ordered 2011-11-02T09:24:16 thestinger: tyvm, is that what you'd recommend doing? 2011-11-02T09:24:28 yeah, no reason not to use C++11 now that you can 2011-11-02T09:24:46 *** Saulzar has joined #aichallenge 2011-11-02T09:25:02 using that data structure like, or would it be easier to just use a 2d int array with different values for explored/not/occupied? 2011-11-02T09:25:20 *** dwins has joined #aichallenge 2011-11-02T09:25:24 2d int array/vector is better for that stuff 2011-11-02T09:25:49 cool, I think that's what I'll use so 2011-11-02T09:25:51 ty 2011-11-02T09:25:55 an unordered_set would be good for how the tutorial deals with unseen locations (you make it once, then remove them) 2011-11-02T09:26:51 caution: every turn is 2 phase, move, then battle resolution 2011-11-02T09:26:59 hmmm, ok. I'll follow along with the tutorial as closely as I can for now so. They mention that it's used further along too so it might to be my advantage to learn now 2011-11-02T09:27:04 food resolutionshould go after battle, i think. 2011-11-02T09:27:24 bobbydroptable: this is just for avoiding my own ants 2011-11-02T09:27:45 thanks thestinger 2011-11-02T09:28:14 conor_f my message was for Caution. lol 2011-11-02T09:28:28 following the tutorial will actually give you a pretty good bot to start from, in python it's too slow and has timeouts but it's fine in C++ 2011-11-02T09:28:57 bobbydroptable: my bad :P 2011-11-02T09:29:49 *** Saulzar has quit IRC (Ping timeout: 252 seconds) 2011-11-02T09:29:53 thestinger: ok, I suppose its good for me to start using new data structures and making my code more C++ then C like I usually do 2011-11-02T09:30:42 is it just me or do the forum posts not work anymore 2011-11-02T09:30:58 it's like post requests don't get through 2011-11-02T09:31:29 *** argiopeweb has quit IRC (Ping timeout: 265 seconds) 2011-11-02T09:31:58 *** formallyknownasd has quit IRC (Ping timeout: 265 seconds) 2011-11-02T09:33:22 are we allowed to include librarys in C++ or do we have to stick to std C++? 2011-11-02T09:34:57 *** patrisk has quit IRC (Remote host closed the connection) 2011-11-02T09:35:50 *** zerofoxdodo has joined #aichallenge 2011-11-02T09:36:07 conor_f: if it compiles, sure 2011-11-02T09:36:22 conor_f: they might not exist on the workers, and even if they do, they might not be linked 2011-11-02T09:36:36 k, ty 2011-11-02T09:36:40 <^5> "Next game should be within -1.5 minutes." 2011-11-02T09:36:48 conor_f: so any kind of misc library you include you have to manually include it 2011-11-02T09:37:07 ^5: remember the windows progress estimates? think of that 2011-11-02T09:38:18 <^5> Ah, lol. Its now -1.2, im not sure if thats better or worse 2011-11-02T09:39:42 lol 2011-11-02T09:40:35 *** adhlssu07 has joined #aichallenge 2011-11-02T09:45:08 *** adhlssu07 has left #aichallenge 2011-11-02T09:45:17 *** Areks_ has joined #aichallenge 2011-11-02T09:45:43 *** cafaro has joined #aichallenge 2011-11-02T09:46:52 hey conor im also doing stanford intro to ai 2011-11-02T09:47:21 *** svujic has joined #aichallenge 2011-11-02T09:47:34 i have to agree with you.. he is going way too fast especially when he wants to prove a new function 2011-11-02T09:47:58 i can mechanically apply his math and all but trying to understand it intuitively is hard sometimes 2011-11-02T09:48:15 *** marijnfs has quit IRC (Quit: leaving) 2011-11-02T09:48:31 *** marijnfs has joined #aichallenge 2011-11-02T09:48:35 and with other parts i think "sure i understand it now but in two days ill have forgotten this" 2011-11-02T09:51:38 yeah, it could have been made a lot better imo :/ 2011-11-02T09:51:50 if you compare it to kahn's academy for example, kahn is much better, even though the repetition can make things a bit boring 2011-11-02T09:51:52 *** pedrosorio has joined #aichallenge 2011-11-02T09:52:03 urgh, putting rotating bias into exploration routine seems like an ugly hack 2011-11-02T09:52:57 and doesn't work very well, now that i see it 2011-11-02T09:52:59 ==" 2011-11-02T09:55:46 *** TTE has joined #aichallenge 2011-11-02T09:56:36 bobby: why doesn't it work well? it should evenly spread out the ants over all directions? 2011-11-02T09:57:35 *** Rinum has quit IRC (Ping timeout: 265 seconds) 2011-11-02T09:57:43 *** rajanaresh has joined #aichallenge 2011-11-02T10:03:59 *** mviel_ has quit IRC (Ping timeout: 248 seconds) 2011-11-02T10:04:18 bad heuristics i suppose 2011-11-02T10:04:23 *** mviel_ has joined #aichallenge 2011-11-02T10:04:52 *** pedrosorio has quit IRC (Quit: Page closed) 2011-11-02T10:05:00 *** Harpyon has joined #aichallenge 2011-11-02T10:08:51 *** mleise has quit IRC (Quit: Leaving.) 2011-11-02T10:09:24 *** xathis has quit IRC (Quit: Page closed) 2011-11-02T10:13:45 *** thestinger has quit IRC (Quit: WeeChat 0.3.6) 2011-11-02T10:15:59 *** treeform has joined #aichallenge 2011-11-02T10:16:20 *** AxVapor has quit IRC (Quit: AxVapor) 2011-11-02T10:18:53 *** conor_f has quit IRC (Ping timeout: 256 seconds) 2011-11-02T10:21:59 *** Antimony has joined #aichallenge 2011-11-02T10:23:33 *** retybok has joined #aichallenge 2011-11-02T10:23:52 does someone know how to use WriteDebugImage in the Go package? 2011-11-02T10:24:19 I used it although I had to change it I think. 2011-11-02T10:25:40 *** balkemacho has quit IRC (Quit: Leaving) 2011-11-02T10:28:50 *** ChrisH has joined #aichallenge 2011-11-02T10:29:32 *** g0llum has joined #aichallenge 2011-11-02T10:30:59 unf. it now includes a bunch of my secret sauce...what problem did you run into? 2011-11-02T10:35:22 how to supply that function that returns colors for a position 2011-11-02T10:35:47 but apparently you can define teh function inline and it can have access to any variables at that scope 2011-11-02T10:35:51 didnt know that 2011-11-02T10:36:09 m.WriteDebugImage("_test", 0, func(c, r int) image.NRGBAColor { return m.At(r, c) }) 2011-11-02T10:37:08 *** bmh has joined #aichallenge 2011-11-02T10:37:21 was the arg how I called it. and iirc I changed func (* Map)At to return image.NRGBAColor. 2011-11-02T10:38:48 @later tell janzert Commenting out the rate limitting code may produce better parings, but it seems to have a bias towards submissions with more total games played (which is not surprising because they provide more information), but those submissions aren't necessarily active, so that doesn't help the deactivation policy. 2011-11-02T10:38:48 ChrisH: OK 2011-11-02T10:39:36 ChrisH might know better than me since I am 2 weeks out from hello.go 2011-11-02T10:39:44 :) 2011-11-02T10:39:47 *** olexs has quit IRC (Quit: Leaving.) 2011-11-02T10:39:53 @later tell janzert What about keeping the rate limiting code, but allowing for a higher rate for active submissions? 2011-11-02T10:39:53 ChrisH: Job's done. 2011-11-02T10:41:09 I don't use the code from the go starter bot because it didn't compile when I started my bot code back in beta, so I'm not sure how much help I'll be with the graphics stuff in that code. 2011-11-02T10:41:37 yeah you are right jcdny 2011-11-02T10:42:09 i didnt know such functions could use local variables, like 'm' and keep that access 2011-11-02T10:42:28 but m.At is a Color 2011-11-02T10:42:32 *** scombinator has joined #aichallenge 2011-11-02T10:42:36 not a NRGBAColor 2011-11-02T10:42:40 it whines about that 2011-11-02T10:42:45 I'm here now, you can all talk about your strategies 2011-11-02T10:42:57 so i got the item and used its color function 2011-11-02T10:42:59 *** ChrisH has quit IRC (Quit: Page closed) 2011-11-02T10:43:02 which returns nrgbacolor 2011-11-02T10:43:23 what EC2 instances is this running on? 2011-11-02T10:43:30 *** Overburn has quit IRC (Quit: Leaving) 2011-11-02T10:43:39 *** bmh has quit IRC (Quit: bmh) 2011-11-02T10:44:13 *** Areks_ has quit IRC (Ping timeout: 240 seconds) 2011-11-02T10:44:14 ChrisH yeah I binned most of the starter though I did keep the image bit for debugging stuff which has been fun. I layer on various things like influence maps etc to see how the heuristics are working 2011-11-02T10:44:33 marijnfs: those functions are called closures, it's one of Go's main features (google it) 2011-11-02T10:45:20 marijnfs: I can't help you about this particular function (I don't use it), but I'm sure you will sort it out. I've been using the Go starter package for a while now and it's been very productive. 2011-11-02T10:45:47 yeah i got it, it makes pictures now 2011-11-02T10:45:55 i now get the hang of Go;) 2011-11-02T10:45:56 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-11-02T10:46:04 but it takes some getting used to 2011-11-02T10:46:12 the error messages are good now 2011-11-02T10:46:23 marijnfs: sure, it took me some time to (it was my first experiece with Go) 2011-11-02T10:46:23 i used it before and just quit because of weird messages 2011-11-02T10:46:46 today i played with the heap datastructure 2011-11-02T10:46:51 it took some playing around 2011-11-02T10:47:10 because you have to create your own class and then operate on it 2011-11-02T10:47:19 but finally it worked 2011-11-02T10:47:36 i just wish it had generics or something 2011-11-02T10:47:51 marijnfs: what underlying container did you use? 2011-11-02T10:48:05 generics would be nice, but I think interfaces are good enough 2011-11-02T10:48:18 *** AxVapor has joined #aichallenge 2011-11-02T10:49:47 *** amstan has quit IRC (Ping timeout: 258 seconds) 2011-11-02T10:50:44 yeah the stuff i had to do for heap were to do what you normally do with generics 2011-11-02T10:50:52 its not too much of a problem 2011-11-02T10:51:11 but it involves these empty interfaces and runtime conversion 2011-11-02T10:51:14 kind of nasty 2011-11-02T10:54:26 the go starter package keeps track of water once its seen it 2011-11-02T10:54:36 but seen land becomes unknown again right? 2011-11-02T10:55:10 marijnfs: yes, I think so. I think it makes sense, no? 2011-11-02T10:55:18 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-02T10:55:23 Actually I have to read that code again :) 2011-11-02T10:56:42 *** olexs has joined #aichallenge 2011-11-02T10:58:23 im guessing this from the pictures it draws 2011-11-02T10:58:30 there is water standing 2011-11-02T10:58:38 but the ground around it is gray again 2011-11-02T10:58:38 *** AxVapor has quit IRC (Quit: AxVapor) 2011-11-02T10:59:16 i want to create a frontier for exploration so i need that land to stay 2011-11-02T10:59:22 but ill just hack it in the map somehwere 2011-11-02T10:59:39 *** conor_f has joined #aichallenge 2011-11-02T10:59:47 *** AxVapor has joined #aichallenge 2011-11-02T11:01:32 marijnfs: I don't hesitate to add fields in the map :) 2011-11-02T11:02:23 yeah thats the easiest i guess 2011-11-02T11:02:32 i also want ants to have some state variables 2011-11-02T11:02:37 *** Antimony has joined #aichallenge 2011-11-02T11:02:50 at the moment it is hard to track them 2011-11-02T11:02:53 how would you do that 2011-11-02T11:06:47 marijnfs: I guess the only way is to track the orders you give them 2011-11-02T11:06:56 why do you want to do that though? 2011-11-02T11:08:08 ow stuff like fighters and defenders 2011-11-02T11:08:27 or give them some memory 2011-11-02T11:08:48 *** Harpyon_ has joined #aichallenge 2011-11-02T11:10:35 *** Pawel__ has joined #aichallenge 2011-11-02T11:10:59 I track all my ants so I can store the paths from frame to frame 2011-11-02T11:10:59 otherwise it is hard to make them divide tasks 2011-11-02T11:11:03 and avoid recalculating 2011-11-02T11:11:29 each ant has a path it tries to follow, and a goaltype 2011-11-02T11:11:45 ok you store the planned paths? 2011-11-02T11:12:06 if the type of square at the end of the path changes, they erase their path and get assigned a new oe 2011-11-02T11:12:31 Hello - I have test my bot with a script test_bot from starter package. Everything was ok and now when I submitted - I have received information that my bot didn't compile and I have 32 ERRORS... What is wrong with this competition?? 2011-11-02T11:13:21 nnn 2011-11-02T11:13:26 whats wrong with your code 2011-11-02T11:13:37 you probably didn't send it properly 2011-11-02T11:13:47 was the binary still in the dir 2011-11-02T11:13:56 How there can be anything wrong with my code if it runs on my machine with the test script?? 2011-11-02T11:13:58 or you use libraries that they dont have 2011-11-02T11:14:15 *** sigh has quit IRC (Remote host closed the connection) 2011-11-02T11:14:18 I am using java 7# 2011-11-02T11:14:25 compiling doesnt have to do with the test script 2011-11-02T11:14:34 java 7 - could this be a problem 2011-11-02T11:14:46 ? 2011-11-02T11:14:50 i guess if they use 6 2011-11-02T11:14:55 and you use new features 2011-11-02T11:14:58 dont know what they use 2011-11-02T11:15:23 maybe you should somehow explicitly say you use java 7 in the makefile somewhere 2011-11-02T11:15:30 Is there a Location struct for the ants current location? 2011-11-02T11:15:38 yes 2011-11-02T11:15:40 but im not in the Java-scene 2011-11-02T11:15:56 Location.h 2011-11-02T11:16:03 Ok - thanks for this maybe - I will try to use java 6 - but they are similar any way 2011-11-02T11:16:18 so state.myAnts.location? 2011-11-02T11:16:24 or something similar? 2011-11-02T11:16:27 but it would be nice if they send the errors from compilation 2011-11-02T11:16:33 just look at the source files in the starter bot 2011-11-02T11:16:41 OpenJDK 6 (IcedTea6 1.10.2) (6b22-1.10.2-0ubuntu1~11.04.1) 2011-11-02T11:16:41 they're pretty simple 2011-11-02T11:16:47 yeah I was Antimony, but I never looked in state.h 2011-11-02T11:16:47 from http://aichallenge.org/starter_packages.php 2011-11-02T11:16:49 my bad 2011-11-02T11:16:59 I did - and technically there wa nothing wrong 2011-11-02T11:18:01 Does anyone try to track enemy ants? 2011-11-02T11:18:18 The problem is there's no way to see their assigned moves, so it's much harder to track 2011-11-02T11:19:19 *** Pawel__ has quit IRC (Quit: Page closed) 2011-11-02T11:19:27 Antimony: Identity is overrated. 2011-11-02T11:19:46 Antimony: Heck, if you're awesome enough, if two ants meet head-on and want to swap places, you can swap brains and get a free move :D 2011-11-02T11:20:03 i think just reacting to enemies when you see them is enough for now 2011-11-02T11:20:32 In the end it doesn't matter how an ant got somewhere, just that it is here now 2011-11-02T11:20:46 the only reason you might want to do this is efficiency 2011-11-02T11:21:11 Antimony, and what's the use of tracking them? Knowing that one is a coward and the other is an aggressive maniac? :P 2011-11-02T11:21:52 Migi32: it would be nice to focus on the maniac :) 2011-11-02T11:22:05 *** aerique has quit IRC (Quit: ...) 2011-11-02T11:22:07 *** JanneP has joined #aichallenge 2011-11-02T11:22:44 unfortunately I know of no bots that give their ants their own personality 2011-11-02T11:22:58 it would be awesome though :) 2011-11-02T11:23:35 i am thinking of having aggressive ants and workers 2011-11-02T11:23:42 that would be cool 2011-11-02T11:23:51 *** ikaros has joined #aichallenge 2011-11-02T11:23:56 and give them confidence if they are in a group 2011-11-02T11:24:06 that's not a bad idea, but why make it static? 2011-11-02T11:24:18 *** rajanaresh has left #aichallenge 2011-11-02T11:25:23 *** Knekkebjoern has joined #aichallenge 2011-11-02T11:25:27 I have hippie ants and patriot ants 2011-11-02T11:25:39 and sometimes I redistribute all my ants between the two groups 2011-11-02T11:26:35 *** AxVapor has quit IRC (Quit: AxVapor) 2011-11-02T11:26:55 if I have less than 30 ants, they're all hippies 2011-11-02T11:27:00 the hippie ants are trying to be friends with the enemy? 2011-11-02T11:27:41 no but they're less aggressive 2011-11-02T11:27:47 i see your hippie ants are avoiding enemies 2011-11-02T11:28:17 *** nickliti has joined #aichallenge 2011-11-02T11:28:23 do your hippie ants get stoned when they eat food? 2011-11-02T11:28:34 you do that before you eat food 2011-11-02T11:28:39 *** oddmunds has left #aichallenge 2011-11-02T11:28:59 *** UncleVasya has quit IRC (Ping timeout: 244 seconds) 2011-11-02T11:29:14 *** Relax has joined #aichallenge 2011-11-02T11:29:40 *** nickliti has quit IRC (Client Quit) 2011-11-02T11:29:41 *** magiik has quit IRC (Remote host closed the connection) 2011-11-02T11:30:04 *** delt0r__ has quit IRC (Ping timeout: 258 seconds) 2011-11-02T11:30:54 *** JamesMG_ has joined #aichallenge 2011-11-02T11:31:01 *** JamesMG has quit IRC (Quit: leaving) 2011-11-02T11:32:05 *** JamesMG_ has quit IRC (Client Quit) 2011-11-02T11:32:29 *** JamesMG has joined #aichallenge 2011-11-02T11:34:08 *** UncleVasya has joined #aichallenge 2011-11-02T11:35:57 *** smiley1983 has quit IRC (Quit: leaving) 2011-11-02T11:38:16 *** Larose has joined #aichallenge 2011-11-02T11:38:38 *** bobbydroptable has quit IRC (Quit: Leaving) 2011-11-02T11:38:38 does food block ants somehow? 2011-11-02T11:39:27 I think I just found out you can't move on top of food 2011-11-02T11:39:34 ? 2011-11-02T11:39:40 *** lavalamp has joined #aichallenge 2011-11-02T11:40:02 you eat the food before you can move on it right? 2011-11-02T11:40:28 if it spawns next to you, you can't move on top of it to eat it 2011-11-02T11:40:47 aha 2011-11-02T11:42:18 *** delt0r__ has joined #aichallenge 2011-11-02T11:43:37 would someone who knows some C++ mind answering a few questions for me on sets? 2011-11-02T11:43:51 ahaa... that might explain the unexplained bumps between my own ants, as I'm not checking if there is food blocking the move.. 2011-11-02T11:43:53 PM would be better so not to clog the channel 2011-11-02T11:44:06 *** treeform has quit IRC (Quit: Leaving.) 2011-11-02T11:44:45 #C++ 2011-11-02T11:45:02 *** Mooloo has joined #aichallenge 2011-11-02T11:46:18 I can't tell if my ants are eating food on purpose or by accident 2011-11-02T11:46:53 *** Relax has quit IRC (Quit: Page closed) 2011-11-02T11:46:56 *** jemhoff_ has joined #aichallenge 2011-11-02T11:47:00 *** jemhoff_ is now known as Relax 2011-11-02T11:47:06 what questions do you have? 2011-11-02T11:47:31 *** halligalli has joined #aichallenge 2011-11-02T11:47:41 And yes, that was the reason that I wanted to track enemy ants. So I could guess patterns in their behavior 2011-11-02T11:47:44 Antimony: just some strange errors and it's probably because I'm doing something stupid with sets as I've never used them before 2011-11-02T11:47:48 *** conor_f has quit IRC (Quit: leaving) 2011-11-02T11:47:51 what error? 2011-11-02T11:48:16 I figure ants tend to make the same move several times in a row, or oscillate between two moves 2011-11-02T11:48:21 *** conor_f has joined #aichallenge 2011-11-02T11:48:25 *** Baus has quit IRC (Ping timeout: 256 seconds) 2011-11-02T11:49:05 meh, which bot is this one? http://aichallenge.org/visualizer.php?game=48863 2011-11-02T11:49:21 *** pairofdice has joined #aichallenge 2011-11-02T11:50:07 if(destLocations.find(loc) != destLocations.end()) 2011-11-02T11:50:11 that line for example 2011-11-02T11:50:16 shuld it be posible to run multiple games with playgame.py? 2011-11-02T11:50:31 destLocations is defined as set destLocations; 2011-11-02T11:50:44 what is the error? 2011-11-02T11:51:22 by the way, if you just want to know if an element is in the set, use count(x) 2011-11-02T11:51:27 oh 2011-11-02T11:51:28 ok 2011-11-02T11:52:36 *** Hexren has joined #aichallenge 2011-11-02T11:52:37 hmm 2011-11-02T11:53:35 *** Antimony_ has joined #aichallenge 2011-11-02T11:56:08 *** Antimony has quit IRC (Ping timeout: 276 seconds) 2011-11-02T11:57:05 http://pastebin.com/qY4Dxjtf 2011-11-02T11:57:21 thats the error when I use if(destLocations.count(loc)) 2011-11-02T11:57:29 @later tell janzert that shouldn't have fought against deactivation, because the average remained the average of only the active bots (line 150) 2011-11-02T11:57:29 McLeopold: Ready to serve, my lord. 2011-11-02T11:58:01 *** Antimony_ has quit IRC (Ping timeout: 258 seconds) 2011-11-02T12:00:12 *** marijnfs has quit IRC (Quit: leaving) 2011-11-02T12:00:20 *** Antimony has joined #aichallenge 2011-11-02T12:00:32 w 2011-11-02T12:00:34 *wb 2011-11-02T12:00:43 15:56 < conor_f> http://pastebin.com/qY4Dxjtf 2011-11-02T12:00:43 15:56 < conor_f> thats the error when I use if(destLocations.count(loc)) 2011-11-02T12:01:26 *** muha has joined #aichallenge 2011-11-02T12:03:40 *** simplex_ has joined #aichallenge 2011-11-02T12:04:35 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-11-02T12:05:05 conor_f: Does Location have bool operator<(Location const & other) const { ... } ? 2011-11-02T12:05:13 *** ikaros has joined #aichallenge 2011-11-02T12:06:07 nope 2011-11-02T12:06:21 just two vars, r and c (row and column ints) 2011-11-02T12:06:51 I think this is why you get this error 2011-11-02T12:07:26 ok, why and how to fix? 2011-11-02T12:07:46 you need to define a comparison operator for Location 2011-11-02T12:07:57 *** onensora has quit IRC () 2011-11-02T12:08:05 Because std::set is an ordered set 2011-11-02T12:08:16 hey, does anyone know how the --scenario option of playgame.py works ? 2011-11-02T12:08:18 here's mine: static bool operator<(const Location &a, const Location &b) { return a.row == b.row ? a.col < b.col : a.row < b.row; } 2011-11-02T12:08:43 *** AxVapor has joined #aichallenge 2011-11-02T12:08:48 liberforce: it keeps ants and food on the map and doesn't create ants at the hills 2011-11-02T12:09:11 *** Chirmaya has joined #aichallenge 2011-11-02T12:09:42 McLeopold: i'm not sure i'm getting it... 2011-11-02T12:09:53 a1k0n: ok, so I need to make it unordered or I need to have a bool function that orders them? 2011-11-02T12:10:16 conor_f: yes. just add the line i gave you there to Location.h and set will work. 2011-11-02T12:10:27 McLeopold: how do ants spawn then ? 2011-11-02T12:11:09 they would still spawn normally, it's just that initial ants on a map aren't replaced by 1 ant on each hill 2011-11-02T12:12:08 a1k0n: ok 2011-11-02T12:12:30 *** zerofoxdodo has quit IRC (Remote host closed the connection) 2011-11-02T12:12:34 Hey there. Now that I have some basic functionality with my ants, my current biggest problem is my ants getting stuck in mazes around my hive. Are there some good strategies that I can look at for this? 2011-11-02T12:12:42 *** george_ has joined #aichallenge 2011-11-02T12:13:00 A* 2011-11-02T12:13:30 *** Antimony_ has joined #aichallenge 2011-11-02T12:13:45 I use the A* pathing, but I was more referring to a strategy to use for the AI to move ants far enough away from the hive 2011-11-02T12:13:46 http://en.wikipedia.org/wiki/A*_search_algorithm 2011-11-02T12:13:55 McLeopold: thanks, I think i'm going to try. I first thought that this was the option to replay the server input that can be exported from played games 2011-11-02T12:13:57 if you find an area is a dead end, you can mark all the squares so future a* searches don't go in that area, unless the target square is in there 2011-11-02T12:14:06 *** amstan has joined #aichallenge 2011-11-02T12:14:06 *** ChanServ sets mode: +o amstan 2011-11-02T12:14:14 some people tried making ants repelent 2011-11-02T12:14:23 so that your ants run away from each other 2011-11-02T12:15:28 anyone know if input to bot includes dead ants that are now out of sight - eg my lone ant is killed, so i no longer see that area - but it is still reported, no? 2011-11-02T12:15:48 it should still be reported 2011-11-02T12:16:07 a1k0n: tyvm, working :) 2011-11-02T12:16:10 ok, but only for my own ant - not an enemy ant that died in the combat also? 2011-11-02T12:16:25 conor_f: np 2011-11-02T12:16:27 *** Antimony has quit IRC (Ping timeout: 258 seconds) 2011-11-02T12:16:39 *** Antimony_ is now known as Antimony 2011-11-02T12:16:58 *** Knekkebjoern1 has joined #aichallenge 2011-11-02T12:17:09 *** SITZ has joined #aichallenge 2011-11-02T12:17:24 *** SITZ has left #aichallenge 2011-11-02T12:17:27 @McLeopold nope I dont think so 2011-11-02T12:17:28 Information is given for ants that died during the collision or battle resolution of the previous turn if it is in a square currently visible by one of your live ants. These squares are merely for your information if you wish to use them, they can otherwise be thought of as land and moved into that turn. 2011-11-02T12:17:29 Hexren: No! 2011-11-02T12:17:54 Hexren - it does indeed say that 2011-11-02T12:18:06 BUT in fact the dead ants are reported 2011-11-02T12:18:17 it seems , if they are your own 2011-11-02T12:18:23 ok 2011-11-02T12:18:35 I havent checked the actuakll input yet 2011-11-02T12:18:58 McLeopold - what is the correct official version, do you know, please? 2011-11-02T12:19:14 Why so many people try to use '@' while talking to someone? 2011-11-02T12:19:23 twitter 2011-11-02T12:19:36 Hexren - you can try eg viewing a game from the contest - look at bot input (from the button for bot input) 2011-11-02T12:19:45 irssi tab completion ftw caution :P 2011-11-02T12:20:11 *** Knekkebjoern has quit IRC (Ping timeout: 276 seconds) 2011-11-02T12:20:43 UncleV @@@ i dont know. long time now see. hows things? 2011-11-02T12:20:54 (and do you know the answer to my question?) 2011-11-02T12:22:03 you get the dead ant info, but not the dead ants vision 2011-11-02T12:22:29 ok, thanks. 2011-11-02T12:22:37 if it is because of a collision, you won't know because you won't get 2 dead ant lines at the same square, just yours 2011-11-02T12:23:52 maybe edit the webpage 'specification' at the end of the bot input section, to clarify - currently it says what Hexren quoted, which is not (as you tell me) correct. 2011-11-02T12:23:57 thanks for that. :) 2011-11-02T12:24:40 *** Harpyon_ has quit IRC (Quit: Computer has gone to sleep.) 2011-11-02T12:25:25 *** conor_f has quit IRC (Ping timeout: 255 seconds) 2011-11-02T12:28:01 *** dwins has quit IRC (Ping timeout: 252 seconds) 2011-11-02T12:28:56 Hexren: where did you find that? 2011-11-02T12:29:10 *** pedrosorio has joined #aichallenge 2011-11-02T12:30:31 http://aichallenge.org/specification.php 2011-11-02T12:30:32 Hexren, you will see your own dead ants, even if they're your only ant in the area 2011-11-02T12:30:50 at the end of the "bot input" sectrion 2011-11-02T12:31:13 i checked 2011-11-02T12:31:31 "I see dead ants"... 2011-11-02T12:31:35 spooky 2011-11-02T12:31:39 :) 2011-11-02T12:31:59 *** dwins has joined #aichallenge 2011-11-02T12:32:05 cyphase - you are correct. but the specs say differently (incorrect;y apparently) 2011-11-02T12:32:56 *** Overburn has joined #aichallenge 2011-11-02T12:33:24 george_: ok, updated 2011-11-02T12:33:24 *** CrazyMLC has joined #aichallenge 2011-11-02T12:33:36 I come in peace 2011-11-02T12:33:55 *** Chirmaya has quit IRC (Quit: Page closed) 2011-11-02T12:33:56 you go in pieces 2011-11-02T12:33:57 McLeopold. thnaks for that. 2011-11-02T12:34:07 AXEMURDERER 2011-11-02T12:34:23 it's updated already.. nice 2011-11-02T12:34:38 hm if that is wanted i wouldnt need to track my ants anymore from turn to turn 2011-11-02T12:34:39 wtf anyone got any idea why the launcher gives me failed to start? 2011-11-02T12:36:03 CrazyMLC: http://www.youtube.com/watch?v=I8F92p1ept0 2011-11-02T12:36:09 what a terrible movie 2011-11-02T12:36:25 so if two own ants go to one tile and so both die are there 2 dead ants sent in the next turn or only 1 2011-11-02T12:36:56 the specific tile of the map will be d 2011-11-02T12:37:10 k 2011-11-02T12:37:40 so you won't know that 2 died unless you do checks 2011-11-02T12:37:53 *** scombinator has quit IRC (Remote host closed the connection) 2011-11-02T12:38:03 *** FastEddie_ has quit IRC (Quit: WeeChat 0.2.6) 2011-11-02T12:38:04 i think i just keep the old system 2011-11-02T12:38:18 what do you mean? 2011-11-02T12:38:21 2 dead ants are sent 2011-11-02T12:38:38 wait, i must have gotten the questions wrong 2011-11-02T12:38:44 what exactly does sent mean in this context? 2011-11-02T12:38:54 how i track my ants 2011-11-02T12:39:09 some of the starter packs only use a matrix, so you can't stuff 2 values into a square, but 2 lines would have been sent 2011-11-02T12:39:21 hmm i get one line only 2011-11-02T12:39:26 as far as i remember 2011-11-02T12:39:27 lemme check 2011-11-02T12:40:44 ah no 2011-11-02T12:40:46 my bad 2011-11-02T12:40:53 i get one line for each ant that dies 2011-11-02T12:41:01 hmm why didn't i notice that before 2011-11-02T12:41:03 useful stuff 2011-11-02T12:42:29 *** halligalli has quit IRC (Quit: Page closed) 2011-11-02T12:42:31 McLeopold: wrap checkbox input and "Remember me" with