2011-12-01T00:00:26 probably so it can say dead[owner]++ 2011-12-01T00:01:01 hm 2011-12-01T00:01:04 thanks 2011-12-01T00:03:00 *** amstan has quit IRC (Read error: Connection reset by peer) 2011-12-01T00:03:06 *** amstan_ has joined #aichallenge 2011-12-01T00:03:06 *** ChanServ sets mode: +o amstan_ 2011-12-01T00:04:17 http://ants.fluxid.pl/replay.7050 2011-12-01T00:04:52 I found that game amusing 2011-12-01T00:05:10 something definitely wrong with my ants there 2011-12-01T00:05:22 strcat_dense in an ant sandwich 2011-12-01T00:06:22 That's dense alright 2011-12-01T00:06:30 I think I've got to throw away my movement code again 2011-12-01T00:06:54 Those huge plugs don't look very efficient 2011-12-01T00:09:07 *** delt0r_ has joined #aichallenge 2011-12-01T00:09:08 *** rajanaresh has joined #aichallenge 2011-12-01T00:10:57 *** yoden has quit IRC (Quit: Leaving.) 2011-12-01T00:12:34 Damn my internet connection.. keep timing out on fluxid.. trying tcpants... 2011-12-01T00:15:24 so what happened to the servers? 2011-12-01T00:15:29 *** rajanaresh has quit IRC (Ping timeout: 245 seconds) 2011-12-01T00:18:22 *** ThiagoRRamos has joined #aichallenge 2011-12-01T00:23:05 *** Antimony has quit IRC (Ping timeout: 252 seconds) 2011-12-01T00:33:41 *** Antimony has joined #aichallenge 2011-12-01T00:47:25 *** Seeker` has joined #aichallenge 2011-12-01T00:47:33 *** ThiagoRRamos has quit IRC (Ping timeout: 260 seconds) 2011-12-01T00:51:54 BenJackson: what's the goal, try to minimize the points other bots get by not letting the cuttoff rules kick in? 2011-12-01T00:52:49 well I was mostly thinking "because I can" 2011-12-01T00:52:58 but yes, it seems odd to me that guy got "2nd" 2011-12-01T00:53:12 how long will it take to code? 2011-12-01T00:53:12 it doesn't reflect anything meaningful about that bot's skill 2011-12-01T00:53:43 *** antimatroid has quit IRC (Ping timeout: 244 seconds) 2011-12-01T00:53:43 ChrisH_: Just to synchronize two kills? Probably not much 2011-12-01T00:53:59 to avoid accidentally causing that case to lose hills or games ... a lot more 2011-12-01T00:54:24 it's an utter fluke in this case the hill caps were only 1 turn apart 2011-12-01T00:54:46 then it sounds like the time spent avoiding the risk of it making your bot worse outweighs any advantage you gain, so my answer is, no, don't do it. 2011-12-01T00:55:11 i would hope you have more valuable features to add still 2011-12-01T00:55:21 shit, I knew there was a reason you were climing the ranks: you're applying good software engineering principels, aren't you??? 2011-12-01T00:55:23 *** gcflymoto has joined #aichallenge 2011-12-01T00:55:37 BenJackson: I can't help it, that's my job. ;) 2011-12-01T00:55:47 whereas I admire the "GG" guy :) 2011-12-01T00:56:11 does he draw that in ants? 2011-12-01T00:56:19 yeah, that happened very early 2011-12-01T00:56:38 ahh, yeah, my time is more valuable than that 2011-12-01T00:57:27 *** JorgeB has joined #aichallenge 2011-12-01T00:57:32 my actual feeling on the contest is along the lines of: ahh, now that I wrote the prototype I really know what it's about, and I should start from scratch 2011-12-01T00:57:36 It's cute, but I thrive on the competition and how it forces me to think and strategize, both things I enjoy more than drawing. :P 2011-12-01T00:58:01 at the time that bot was winning AND drawing GG 2011-12-01T00:58:17 *** rohantt has quit IRC (Ping timeout: 260 seconds) 2011-12-01T00:58:25 oh, so he was showing off, that I respect. ;) 2011-12-01T00:58:44 winning with STYLE 2011-12-01T00:58:55 actually I respect it either way, because it's about having fun, and clearly the coder was having fun. 2011-12-01T00:59:15 it's just not what I would enjoy 2011-12-01T00:59:44 I actually found it to be interesting from the point of view that I had no aspect to my bot that would have allowed me to spell out anything 2011-12-01T00:59:50 it got me thinking about what his internals must be like 2011-12-01T01:00:38 hmmm, well if I wanted to add such a feature I would need to add a set of ant targets with the GG formation and then they would flow there based on my path finding 2011-12-01T01:00:41 @seen bmh 2011-12-01T01:00:41 BenJackson: bmh was last seen in #aichallenge 22 hours, 19 minutes, and 9 seconds ago: sweet. Timeout Bot v1 is complete. I'm going to bed. 2011-12-01T01:01:18 at the time if I did that it would have just caused all of my ants to converge in a blob surrounding the outline of "GG" 2011-12-01T01:01:47 well yeah, at that time my bot would have probably done that too, but not any more. 2011-12-01T01:02:13 anyway, I got a combat bug to fix, so I'm off IRC to concentrate. another time all 2011-12-01T01:02:27 bmh just emailed asking me to dumb down my bot 2011-12-01T01:02:31 hmmm 2011-12-01T01:02:46 *** ChrisH_ has quit IRC (Quit: Page closed) 2011-12-01T01:03:13 *** antimatroid has joined #aichallenge 2011-12-01T01:03:43 My bot needs to go the other way, badly. Getting whooped. 2011-12-01T01:03:58 are you on bhickey.net ? 2011-12-01T01:04:11 nope, tcpants nad fluxid 2011-12-01T01:04:14 http://bhickey.net:2080/ 2011-12-01T01:04:20 we're trying to keep some lower level bots up there 2011-12-01T01:04:27 to get people started 2011-12-01T01:04:57 Well this has helped me see where my flaws are at least.. will run a few games on there and see what happens. 2011-12-01T01:16:50 *** bmh has joined #aichallenge 2011-12-01T01:17:41 BenJackson: thanks for toning down your bot. It's actually losing sometimes ;) 2011-12-01T01:18:07 *** ThiagoRRamos has joined #aichallenge 2011-12-01T01:18:12 also, nice BFS animation. 2011-12-01T01:19:50 a1k0n: hm? 2011-12-01T01:20:33 I saw the animation also, Found it useful. 2011-12-01T01:20:35 http://forums.aichallenge.org/viewtopic.php?f=24&t=2010 2011-12-01T01:22:27 *** ltriant has quit IRC (Quit: Computer has gone to sleep) 2011-12-01T01:22:48 oh, that is nice. 2011-12-01T01:24:41 Oh thank god, my diffusion code is finally running like it's supposed to. 2011-12-01T01:26:16 If only my ants could do more than spread out and eat... 2011-12-01T01:26:39 *** Cyndre has joined #aichallenge 2011-12-01T01:26:57 *** roflmao has quit IRC (Quit: Leaving.) 2011-12-01T01:32:16 Is it just me or is Java a bit random about when it accepts 1,0 or 1.0 2011-12-01T01:35:09 *** nplus has joined #aichallenge 2011-12-01T01:35:48 *** bmh has quit IRC (Quit: Page closed) 2011-12-01T01:39:53 gazpachoking: Nice. my bot is acting retarded. I think I'm going to go down the diffusion route also on a new bot. 2011-12-01T01:40:29 so now i switched from using fixed key dict to 2d array, and my code is twice slower :/ oops 2011-12-01T01:41:00 Yeah, I'm liking it. I need to tweak all my constants a bit still, but it's working pretty well. 2011-12-01T01:41:30 how are people doing diffusion? 2011-12-01T01:42:12 *** Fandekasp has joined #aichallenge 2011-12-01T01:42:49 I assign starting values to various goals, and diffusion constants based on what is in the cell, then diffuse for a bunch of iterations, and tell my ants to go towards the highest value 2011-12-01T01:42:51 *** Garf has joined #aichallenge 2011-12-01T01:43:09 *diffusion coefficients rather 2011-12-01T01:43:35 *** antimatroid has quit IRC (Read error: Connection reset by peer) 2011-12-01T01:43:43 *** antimatroid has joined #aichallenge 2011-12-01T01:44:40 Hard to do fast enough in python 2011-12-01T01:45:26 Yeah, not sure how much time I have left after this. Using numpy it's something like 400x faster than the straight python though. 2011-12-01T01:45:31 That is similar to what I'm doing in java, but mine persists through frames. It is ok for my first try but I think I will try it a different way. 2011-12-01T01:46:19 After trying so many things my code is really messy and half commented out.. time for a clean slate I think. :) 2011-12-01T01:46:44 Falcon256: I've started with a clean slate every night for the past 3 days :P 2011-12-01T01:47:19 Oh, nice. Mine has lasted three days so far. hehe 2011-12-01T01:47:22 My code is half commented out too right now, but I like this base the best to build on, I think I won't start over again tomorrow. 2011-12-01T01:48:49 what are the tricks for numpy gazpa? 2011-12-01T01:50:57 You can use arrays, and operate on the whole thing at once, which means the only python loop in the diffusion code is the one to iterate through the steps. 2011-12-01T01:51:06 *** Acy- has left #aichallenge 2011-12-01T01:51:12 Let me get you the link for the page I read... 2011-12-01T01:51:16 thanks 2011-12-01T01:51:46 http://www.timteatro.net/2010/10/29/performance-python-solving-the-2d-diffusion-equation-with-numpy/ 2011-12-01T01:52:11 ah, nice 2011-12-01T01:52:38 It's a bit more complicated and slower when you don't have a constant diffusion coefficient over your whole grid.. 2011-12-01T01:53:03 I think there are probably optimizations that can be made to my code still.. 2011-12-01T01:56:54 Oh I should probably make the goals/coefficient arrays part of the same array, so I can roll them at once.. maybe that will help. 2011-12-01T01:57:38 *** Antimony has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111104165243]) 2011-12-01T01:57:54 thank you! 2011-12-01T01:58:09 No problem. 2011-12-01T01:59:15 *** Murashka has joined #aichallenge 2011-12-01T02:03:30 *** mleise has joined #aichallenge 2011-12-01T02:04:35 "Executing the above code on my office workstation for 500 time-steps with dx=dy=0.01 takes about 88 seconds, averaging 0.1758s per time-step. Considering a compiled C-code implementation would execute all 500 time-steps in less than two-tenths of a second, the Python performance is pretty dismal." 2011-12-01T02:05:29 Why is it so slow 2011-12-01T02:05:33 On my computer, the numpy expression takes less than a quarter-second to execute (compare plain Python at 88s) with an average of 4.70094E-4s per time-step: nearly 400-times faster than the plain Python code. 2011-12-01T02:05:38 Keep reading 2011-12-01T02:05:43 That was the python loop version 2011-12-01T02:05:46 Yeah 2011-12-01T02:05:52 Oh 2011-12-01T02:05:53 Dunno 2011-12-01T02:06:03 But it sucks. 2011-12-01T02:10:11 *** epicmonkey has joined #aichallenge 2011-12-01T02:10:32 *** imam has joined #aichallenge 2011-12-01T02:11:07 *** imam has left #aichallenge 2011-12-01T02:14:55 *** epicmonkey has quit IRC (Ping timeout: 248 seconds) 2011-12-01T02:16:38 *** ThiagoRRamos has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111104165243]) 2011-12-01T02:18:06 *** epicmonkey has joined #aichallenge 2011-12-01T02:23:18 Woah, okay. 110 seconds vs 0.6 seconds 2011-12-01T02:24:10 *** kiv has joined #aichallenge 2011-12-01T02:28:21 *** antimatroid1 has joined #aichallenge 2011-12-01T02:28:21 *** antimatroid has quit IRC (Read error: Connection reset by peer) 2011-12-01T02:32:12 *** lawmax has joined #aichallenge 2011-12-01T02:37:13 pita 2011-12-01T02:37:41 *** CowTipperVirus has quit IRC (Ping timeout: 265 seconds) 2011-12-01T02:38:22 http://pastebin.com/QYSw01ir 2011-12-01T02:38:51 both a's have the same path distance to A but only one can get within the battle radius of A next turn 2011-12-01T02:40:24 Can't neither of them get within it in one move? 2011-12-01T02:40:40 the higher 'a' can 2011-12-01T02:40:45 if it moves up 2011-12-01T02:41:26 I thought this was the battle radius? http://pastebin.com/LkF05q8c 2011-12-01T02:41:52 Oh no, I see 2011-12-01T02:41:54 http://pastebin.com/rtqamqjb 2011-12-01T02:43:15 Yeah, got it 2^2 + 1^2 <= 5 Not sure what I was thinking. 2011-12-01T02:44:39 That's how it is 2011-12-01T03:00:23 pairofdice: regarding python performance: what you have to keep in mind is that it is an interpreted language. if course tight inner loops are going to be slower than C. 2011-12-01T03:08:46 *** bhasker has joined #aichallenge 2011-12-01T03:08:59 anyone around whose brains i can pick for a minute 2011-12-01T03:10:30 Only about 130 people. 2011-12-01T03:11:34 so to do a minmax i create a mini state 2011-12-01T03:12:25 basically i find the minimum bounding square around the ants involved with about a clearance of a couple of rows/cols to reduce errors due to world wrapping etc 2011-12-01T03:12:46 the problem i am having is translating world coordinates to map coordinates to my mini map coordinates 2011-12-01T03:13:21 especially at edges where the square may cross boundaries 2011-12-01T03:13:47 my minimap dimensions all go wrong and it ends up duplicating map data 2011-12-01T03:13:52 which causes minimax to do stupid stuff 2011-12-01T03:13:52 *** ikaros has joined #aichallenge 2011-12-01T03:15:11 *** epicmonkey has quit IRC (Ping timeout: 248 seconds) 2011-12-01T03:17:24 anyone? 2011-12-01T03:21:14 *** antimatroid1 has quit IRC (Read error: Connection reset by peer) 2011-12-01T03:21:25 *** antimatroid has joined #aichallenge 2011-12-01T03:26:32 antimatroid 2011-12-01T03:26:39 sup? 2011-12-01T03:26:48 let me read 2011-12-01T03:26:57 *** mj41 has joined #aichallenge 2011-12-01T03:27:15 k 2011-12-01T03:27:29 i don't do that :P 2011-12-01T03:27:30 basicall23:17 you can make u incremental or just go back and forth between tw 2011-12-01T03:27:32 oops 2011-12-01T03:27:52 i do a mini minimax thing which doesn't really require me to 2011-12-01T03:28:29 hmm well i just thought it will be easier to work with a reduced state so that 2011-12-01T03:28:34 i can restrict the stuff i look at 2011-12-01T03:28:42 bhasker: have you considered trying to just pass the thing the locations and non-water blocked moves of each ant? 2011-12-01T03:28:48 and writing up battle resolution etc. on that? 2011-12-01T03:30:17 yea tried that doesn't do so well 2011-12-01T03:30:32 well i am considering stuff like movetuples 2011-12-01T03:30:40 and don't solve it for individual ants 2011-12-01T03:30:46 i try and solve for groups if possible 2011-12-01T03:31:09 and mostly i don't use it to decide where to go 2011-12-01T03:31:19 once i decide where to go i see if its beneficial using minmax 2011-12-01T03:31:23 if not do something else 2011-12-01T03:31:32 so i use it to evaluate specific moves 2011-12-01T03:35:35 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-12-01T03:36:53 *** Falcon256 has quit IRC (Quit: Page closed) 2011-12-01T03:37:00 *** Blkt has joined #aichallenge 2011-12-01T03:40:21 *** mviel has joined #aichallenge 2011-12-01T03:41:47 good morning everyone 2011-12-01T03:45:19 Good mourning 2011-12-01T03:45:37 *** jon1 has joined #aichallenge 2011-12-01T03:45:58 *** zeroG|2 has joined #aichallenge 2011-12-01T03:46:21 *** [0x13] <[0x13]!298939b5@gateway/web/freenode/ip.41.137.57.181> has joined #aichallenge 2011-12-01T03:46:53 <[0x13]> hi, 2011-12-01T03:48:00 <[0x13]> i want know, its normal to have next game shedule more than 1180 MIN O_O",,, thx 2011-12-01T03:48:37 hi 2011-12-01T03:49:04 Not quite 2011-12-01T03:50:06 ZzZzzzzZZZ 2011-12-01T03:51:24 There are the tcp servers of course 2011-12-01T03:53:15 http://ants.fluxid.pl/ http://tcpants.com/ http://www.bhickey.net:2080/ 2011-12-01T03:53:46 Where your ants can play as fast as your computer lets them 2011-12-01T04:01:01 <[0x13]> is the same? using tcp or waiting games? 2011-12-01T04:02:35 my tcp will be down for a little while 2011-12-01T04:04:32 anyone know what the largest map dimensions are? 2011-12-01T04:04:37 200x200 2011-12-01T04:04:51 ok, nice, thanks. 2011-12-01T04:07:02 *** g0llum has joined #aichallenge 2011-12-01T04:11:05 *** ikaros has joined #aichallenge 2011-12-01T04:11:42 *** JorgeB has quit IRC (Quit: Computer has gone to sleep.) 2011-12-01T04:13:34 *** NoxiaZ^ has joined #aichallenge 2011-12-01T04:18:04 *** Murashka has quit IRC (Quit: Page closed) 2011-12-01T04:20:30 *** grwip has joined #aichallenge 2011-12-01T04:20:33 i want too see 10000 turn game on 1000x1000 2011-12-01T04:21:09 restarted the server 2011-12-01T04:21:15 *** mikewintermute has joined #aichallenge 2011-12-01T04:21:16 multihill maps are back 2011-12-01T04:21:45 *** kiv has quit IRC (Quit: Leaving) 2011-12-01T04:25:34 hi, can you help me. I can't connect to ants.fluxid.pl 2011-12-01T04:25:45 example command : path/python.exe path/tcpclient.py ants.fluxid.pl 2081 "path/php.exe path/MyBot-tcp.php" my_name my_pass 2011-12-01T04:25:55 Read up a few lines 2011-12-01T04:26:46 Oh, right 2011-12-01T04:26:49 telnet ants.fluxid.pl 2081 get error "failure to connect" 2011-12-01T04:29:08 http://ants.fluxid.pl/ - disable please conection limit from one IP or double it 2011-12-01T04:34:07 or it protection from fast reconect? 2011-12-01T04:37:55 *** moises has joined #aichallenge 2011-12-01T04:43:48 uh, what 2011-12-01T04:43:49 *** Rav3nsW0rd has joined #aichallenge 2011-12-01T04:43:53 Accoun: what's your bot name? 2011-12-01T04:44:29 there is no connection limit 2011-12-01T04:45:10 you only can't have multiple connections to same username 2011-12-01T04:45:27 *** praveen_ has joined #aichallenge 2011-12-01T04:47:26 *** september5th has joined #aichallenge 2011-12-01T04:48:15 *** september5th has quit IRC (Client Quit) 2011-12-01T04:50:21 *** antimatroid has quit IRC (Ping timeout: 244 seconds) 2011-12-01T04:51:42 *** antimatroid has joined #aichallenge 2011-12-01T04:54:53 *** tmc2 has joined #aichallenge 2011-12-01T04:57:28 tcpclient.py very bad, it crashed in many sitiations 2011-12-01T04:58:03 if boot deadcycled tcpclient.py deadcycled with it 2011-12-01T04:58:33 if connection error, tcpclient.py crashed 2011-12-01T05:00:29 hmm, pguillory will soon take the #1 spot from xathis 2011-12-01T05:01:13 also, nobody made an attempt at optimal task assignment? 2011-12-01T05:01:28 http://aichallenge.org/visualizer.php?game=149682 2011-12-01T05:01:45 all these bots incorrectly start by going right instead of lef! 2011-12-01T05:01:48 *left 2011-12-01T05:04:15 *** Oygron has joined #aichallenge 2011-12-01T05:04:43 *** Oygron has quit IRC (Client Quit) 2011-12-01T05:11:24 *** gcflymoto has quit IRC (Ping timeout: 265 seconds) 2011-12-01T05:12:40 *** mviel_ has joined #aichallenge 2011-12-01T05:14:13 *** nplus has quit IRC (Ping timeout: 240 seconds) 2011-12-01T05:14:35 *** McLeopold has quit IRC (Quit: Leaving.) 2011-12-01T05:15:07 *** nplus has joined #aichallenge 2011-12-01T05:15:49 *** mviel has quit IRC (Ping timeout: 240 seconds) 2011-12-01T05:16:38 *** zeroG has joined #aichallenge 2011-12-01T05:19:36 *** zeroG|2 has quit IRC (Ping timeout: 260 seconds) 2011-12-01T05:22:39 *** antimatroid has quit IRC (Ping timeout: 244 seconds) 2011-12-01T05:25:20 *** Surya has joined #aichallenge 2011-12-01T05:30:15 *** duurtlang has joined #aichallenge 2011-12-01T05:31:19 I wrote an a* class and edited the tile class, but now i don't have a clue how to get my ants to use my algorithm.. because they don't have a state ? 2011-12-01T05:31:26 can anyone give me some hints or something ? 2011-12-01T05:31:34 *** RandomCoder has joined #aichallenge 2011-12-01T05:32:13 *** ikaros has quit IRC (Ping timeout: 240 seconds) 2011-12-01T05:32:39 *** [0x13] <[0x13]!298939b5@gateway/web/freenode/ip.41.137.57.181> has quit IRC (Ping timeout: 265 seconds) 2011-12-01T05:34:27 *** aerique has joined #aichallenge 2011-12-01T05:36:02 *** praveen_ has quit IRC (Ping timeout: 265 seconds) 2011-12-01T05:37:41 edited the tile class? huh? 2011-12-01T05:37:59 it's not necessary to give your ants state 2011-12-01T05:38:16 I simply optimised my A* algorithm so much that I can recompute each ant's path each turn :) 2011-12-01T05:38:59 but assuming that you want to save the A* results rather than recomputing, yes, you'll probably want to give ants state 2011-12-01T05:39:32 *** antimatroid has joined #aichallenge 2011-12-01T05:42:39 i've added some getters and setters to the tile class for te f g and h values 2011-12-01T05:42:55 ah 2011-12-01T05:43:14 but now i have no idea how to give the paths i generate to ants 2011-12-01T05:43:51 (I'm working together with duurtlang) Do we simply replace the Route class with the path of the A* algo? 2011-12-01T05:43:58 well to give ants states you'll want to map their positions to their state object 2011-12-01T05:44:20 I have no idea what language you're using or how your starter pack works 2011-12-01T05:44:31 Java 2011-12-01T05:44:44 And we are just using the starter Java package from the website 2011-12-01T05:44:50 anyway, so you'll need to calculate where each ant will be next turn (assuming it's not dead) 2011-12-01T05:45:00 hmmyes 2011-12-01T05:45:06 and then delete the states of all ants that no longer exist 2011-12-01T05:45:32 but isn't that already implemented in the starter package? Because the ants walk towards food already now 2011-12-01T05:45:52 I doubt it, it's not in any of the starter packs I've looked at 2011-12-01T05:46:06 + the tutorial btw 2011-12-01T05:46:10 oh 2011-12-01T05:46:19 http://aichallenge.org/ants_tutorial_step_2.php 2011-12-01T05:46:25 well I think all the starter packs have stateless ants 2011-12-01T05:46:31 True 2011-12-01T05:46:44 after doign the tutorial, your ants should just walk towards the nearest food each turn, iirc 2011-12-01T05:46:57 yeah, they do that now 2011-12-01T05:47:50 so as I said, map ant positions to state objects 2011-12-01T05:47:51 So we just have to save the path that is calculated for each ant, and check if where they are going still exists (if it's food) 2011-12-01T05:48:04 yes that would work 2011-12-01T05:48:18 k 2011-12-01T05:48:26 Now we gotta figure out how to code that : 2011-12-01T05:48:27 :D 2011-12-01T05:49:04 Thanks for your help, we're gonna try some stuff now 2011-12-01T05:49:14 *** antimatroid has quit IRC (Ping timeout: 255 seconds) 2011-12-01T05:58:35 I don't think I use any of my ants' stateful features anymore 2011-12-01T05:58:51 I never bothered to implement any 2011-12-01T05:58:56 *** olexs has joined #aichallenge 2011-12-01T05:59:26 I needed it in the beginning, but the algorithms I use now are better and make it unnecessary 2011-12-01T05:59:32 who needs state when you complusively optimise everything to the point that you can recompute it every turn :D 2011-12-01T06:00:23 I haven't done that, but yeah 2011-12-01T06:01:25 if an ant has a 'mission' to get a food, that same ant will usually be assigned to that same food the next turn as well 2011-12-01T06:01:32 so no need to give the ant that mission 2011-12-01T06:02:18 and I found that statefully grouping ants in squads for combat was more a hindrance than a tool 2011-12-01T06:02:21 yes exactly 2011-12-01T06:02:39 since groups can be calculated each turn based on proximity 2011-12-01T06:03:32 is it me, or is this contest being determined by who can make their bot do the least stupid things ratherthan being smart 2011-12-01T06:03:55 for example, not computing paths once and following them even though the circumstances are constantly changing 2011-12-01T06:04:24 strength is often determined by the weakest part 2011-12-01T06:04:43 better to everything half-assed than some things great and others not at all 2011-12-01T06:05:24 I think good combat is most important 2011-12-01T06:05:32 pretty difficult to get right 2011-12-01T06:05:53 then you just need to decide on some macro/econ strategies 2011-12-01T06:06:02 maybe do something clever now and then 2011-12-01T06:07:27 *** AntDroid_ has quit IRC (Ping timeout: 265 seconds) 2011-12-01T06:07:31 I'm starting to think that short-term combat is less important than I thought 2011-12-01T06:08:15 I implemented it near-optimally in my bot and I got like 1.0 mu increase 2011-12-01T06:08:18 in particular, maneuvering well 2011-12-01T06:08:35 implemented combat near optimally? 2011-12-01T06:08:48 I'm guessing that's because it doesn't care what/where it's fighting 2011-12-01T06:08:59 well it seems to me that even the best bots suck at maneuvering 2011-12-01T06:09:19 doing optimal local minimax is difficult if the amount of ants in a fight gets big 2011-12-01T06:10:26 I think people are a little too attached to minimax just because it computes a nash equilibrium in a zero sum game 2011-12-01T06:10:43 I'm "attached" to it because it's locally optimal 2011-12-01T06:10:59 no it's not 2011-12-01T06:11:05 combat isn't a zero sum game 2011-12-01T06:11:28 and minimax is really only applicable if you want to make no assumptions about what the opponent will do 2011-12-01T06:11:34 I'd use expectimax instead 2011-12-01T06:11:50 when fighting a top bot, assuming optimal behavior is reasonable 2011-12-01T06:12:00 (I think that's why the top bots stalemate fights so much) 2011-12-01T06:12:09 just what i was about to say 2011-12-01T06:12:25 they're using a minmax rule 2011-12-01T06:12:25 as for the zero sum issue, that depends on the values you assign to the outcomes 2011-12-01T06:12:54 but as I said, that's only "optimal" in one sense (a very conservative one) 2011-12-01T06:13:41 oh well actually, I doubt they're using minmax 2011-12-01T06:13:58 wait, nvm 2011-12-01T06:28:17 *** olexs has quit IRC (Quit: Leaving.) 2011-12-01T06:43:31 *** RandomCoder has quit IRC (Quit: Page closed) 2011-12-01T06:44:43 *** antimatroid has joined #aichallenge 2011-12-01T06:46:07 *** duurtlang has quit IRC (Ping timeout: 265 seconds) 2011-12-01T06:51:21 http://ants.fluxid.pl/player/fluxid_7b4 bah, it sucks 2011-12-01T06:51:47 *** NoxiaZ^ has quit IRC (Ping timeout: 255 seconds) 2011-12-01T07:03:30 *** UncleVasya has joined #aichallenge 2011-12-01T07:04:23 *** ikaros has joined #aichallenge 2011-12-01T07:08:14 *** Conorach has joined #aichallenge 2011-12-01T07:08:37 *** UncleVasya has quit IRC (Remote host closed the connection) 2011-12-01T07:16:13 *** zeroG has quit IRC (Ping timeout: 240 seconds) 2011-12-01T07:21:12 do I have the self restraint to not fix a performance bug in my A* code? 2011-12-01T07:21:29 I'm hopeless 2011-12-01T07:25:51 *** amstan_ has quit IRC (Ping timeout: 248 seconds) 2011-12-01T07:29:29 *** UncleVasya has joined #aichallenge 2011-12-01T07:33:23 lol: http://bash.org/?947444 2011-12-01T07:37:29 *** NoxiaZ^ has joined #aichallenge 2011-12-01T07:41:38 *** liberforce has joined #aichallenge 2011-12-01T07:47:24 *** u_ has joined #aichallenge 2011-12-01T07:49:08 *** sigh has joined #aichallenge 2011-12-01T07:50:21 *** Anilm3 has joined #aichallenge 2011-12-01T07:50:32 Hi 2011-12-01T07:50:55 Hi. 2011-12-01T07:52:03 Hi uncle 2011-12-01T07:56:19 *** delt0r_ has quit IRC (Ping timeout: 245 seconds) 2011-12-01T08:00:19 well, bye 2011-12-01T08:00:21 Qapla' 2011-12-01T08:00:23 *** Anilm3 has quit IRC (Quit: Lost terminal) 2011-12-01T08:04:39 *** GOKAARGH has joined #aichallenge 2011-12-01T08:04:57 slm 2011-12-01T08:05:05 nbr 2011-12-01T08:08:47 *** delt0r_ has joined #aichallenge 2011-12-01T08:16:03 *** babalui has quit IRC (Quit: Page closed) 2011-12-01T08:19:06 silly main server 2011-12-01T08:22:06 *** GOKAARGH has quit IRC (Quit: Page closed) 2011-12-01T08:29:06 *** TheLinker has joined #aichallenge 2011-12-01T08:29:08 *** Palmik has joined #aichallenge 2011-12-01T08:31:27 *** Fandekasp has quit IRC (Quit: Lost terminal) 2011-12-01T08:36:15 *** HaraKiri has joined #aichallenge 2011-12-01T08:36:25 *** QuirionPT has joined #aichallenge 2011-12-01T08:46:38 *** sigh has quit IRC (Remote host closed the connection) 2011-12-01T08:46:39 *** jcazevedo has joined #aichallenge 2011-12-01T08:46:42 *** mikewintermute has quit IRC (Quit: mikewintermute) 2011-12-01T09:03:33 *** moises has quit IRC (Remote host closed the connection) 2011-12-01T09:09:31 *** avdg has joined #aichallenge 2011-12-01T09:13:59 *** avdg has quit IRC (Ping timeout: 255 seconds) 2011-12-01T09:17:56 http://ants.fluxid.pl/replay.7553 bah... 2011-12-01T09:22:24 finally http://llvm.org/docs/ReleaseNotes.html 2011-12-01T09:22:57 I think clang/llvm 3.0 will actually compile my bot... 2011-12-01T09:24:22 thestinger: where is your bot 2011-12-01T09:24:25 i don't see it on tcp 2011-12-01T09:24:49 oh 2011-12-01T09:24:51 nvm 2011-12-01T09:25:25 *** QuirionPT has quit IRC (Ping timeout: 240 seconds) 2011-12-01T09:25:45 http://ants.fluxid.pl/replay.7630 almost got a point... 2011-12-01T09:27:07 my bot mostly fails bacause it is too offnsive :| 2011-12-01T09:27:21 oh, libstdc++ just needs a patch and it will work with clang 3.0 2011-12-01T09:27:29 and then my bot will compile with clang :) 2011-12-01T09:27:33 defensive* 2011-12-01T09:28:00 clang? why do you want that? can you use it on the official server? 2011-12-01T09:28:11 or do you just want better error message? 2011-12-01T09:28:26 tmc2: sane compiler errors, different warnings than g++, static analysis 2011-12-01T09:28:27 (compile error messages) 2011-12-01T09:28:42 ah yes, static analysis, forgot! 2011-12-01T09:28:50 also autocomplete and error checking in my editor (vim) with clang_complete 2011-12-01T09:28:50 ok, godnight 2011-12-01T09:28:52 *good 2011-12-01T09:28:55 cya :P 2011-12-01T09:29:03 interesting 2011-12-01T09:29:06 *** tmc2 has quit IRC (Quit: Leaving) 2011-12-01T09:30:39 *** Baus has quit IRC (Ping timeout: 244 seconds) 2011-12-01T09:33:28 *** avdg has joined #aichallenge 2011-12-01T09:44:13 *** Baus has joined #aichallenge 2011-12-01T09:49:18 http://ants.fluxid.pl/replay.7652 aaargh, my bot is SO stupid! 2011-12-01T09:51:17 BenJackson: http://bhickey.net:2080/replay.1303 , you should put the bot with attack back on bhickey 2011-12-01T09:51:22 that is just funny 2011-12-01T09:51:47 I wanted to test my attack code, and your bot didn't want anything to do with it 2011-12-01T09:58:22 rwest: play on fluxid.pl too :) 2011-12-01T09:58:35 just call it defends :p 2011-12-01T09:59:27 my company proxy doesn't allow connections to 2080 :( 2011-12-01T09:59:51 *** Baus has quit IRC (Read error: Connection reset by peer) 2011-12-01T09:59:56 meh, I wonder if someone knows how to get the pointer to an element from an iterator 2011-12-01T10:01:10 avdg: I think you can just do &*it 2011-12-01T10:01:13 thestinger: I am, but I get my ass kicked hard there when trying to test 2011-12-01T10:01:48 *** ruuhkis has joined #aichallenge 2011-12-01T10:01:54 Hello :) 2011-12-01T10:02:01 hey 2011-12-01T10:02:05 hows it going 2011-12-01T10:02:12 http://ants.fluxid.pl/player/strcat :) 2011-12-01T10:02:19 thestinger: I'm using &(*it) atm :p 2011-12-01T10:02:27 ah 2011-12-01T10:02:44 not sure why strcat is doing that well, I think the rankings just aren't very precise 2011-12-01T10:02:56 it's the same code as strcat_next was using all of last night 2011-12-01T10:03:28 there we go, it dropped :P 2011-12-01T10:03:28 or the results aren't consistent enough 2011-12-01T10:04:10 thestinger: doing attack code with diffusion takes a lot of tweaking, I need weaker bots to play against 2011-12-01T10:04:20 I have almost got it down though 2011-12-01T10:04:41 I think heh 2011-12-01T10:05:52 *** jon1 has quit IRC (Quit: Leaving.) 2011-12-01T10:06:44 *** TheLinker has quit IRC (Quit: Bye) 2011-12-01T10:07:08 thats impressive 2011-12-01T10:08:04 in doing so, I broke maze navigation 2011-12-01T10:08:05 somehow 2011-12-01T10:09:27 *not sure if knows what diffusion means* 2011-12-01T10:09:47 http://en.wikipedia.org/wiki/Diffusion this? 2011-12-01T10:10:05 it's another name for influence maps 2011-12-01T10:10:16 oh :P 2011-12-01T10:13:32 *** ikaros has quit IRC (Ping timeout: 244 seconds) 2011-12-01T10:14:28 rwest: How many iterations of diffusion do you do? 2011-12-01T10:14:53 I started running into time trouble on larger maps with mine, but those are the ones that I need more iterations to successfully navigate.. 2011-12-01T10:14:55 hey gazpa 2011-12-01T10:15:02 Hey 2011-12-01T10:15:03 *** amstan_ has joined #aichallenge 2011-12-01T10:15:03 *** ChanServ sets mode: +o amstan_ 2011-12-01T10:15:09 thanks for the article, i finished my diffusion 2011-12-01T10:15:28 Cool, working well? 2011-12-01T10:15:30 it was pretty helpful 2011-12-01T10:15:40 *** amstan_ is now known as amstan 2011-12-01T10:15:44 yeah, currently about 200 * 5, taking 200ms 2011-12-01T10:15:55 *** ikaros has joined #aichallenge 2011-12-01T10:15:56 so 1000 maps, which isn't bad 2011-12-01T10:17:23 gazpachoking: I do (rows + cols) / 2 diffusions 2011-12-01T10:17:49 *** LouisMartin has quit IRC (Ping timeout: 265 seconds) 2011-12-01T10:19:30 *** skizo has joined #aichallenge 2011-12-01T10:23:37 *** skizo has quit IRC (Client Quit) 2011-12-01T10:30:32 *** mviel__ has joined #aichallenge 2011-12-01T10:31:46 Kommander: you do multiple maps? 2011-12-01T10:34:08 yes 2011-12-01T10:34:12 *** mviel_ has quit IRC (Ping timeout: 244 seconds) 2011-12-01T10:34:39 food, vision, and 3 battle maps 2011-12-01T10:35:44 Kommander: 200 step diffusion on 5 different maps, is that what you mean? 2011-12-01T10:36:14 You must be doing it way better than me, mine is way slow on 1 map right now.- 2011-12-01T10:36:19 *** jon1 has joined #aichallenge 2011-12-01T10:37:18 well, slightly less 2011-12-01T10:37:26 but close, yeah 2011-12-01T10:37:46 it's more the max(row,col) * 5 2011-12-01T10:38:15 but they are simple maps with mostly zeros 2011-12-01T10:39:47 gazpa: i pretty much did it by the article though, didn't optimize it at all 2011-12-01T10:41:16 Kommander: I combine all my maps 2011-12-01T10:41:31 and only have to diffuse the values over 1 map 2011-12-01T10:42:05 i think i ought to try that, i only have many maps because my last version was way too slow 2011-12-01T10:42:14 rwest: Yeah, that's what I do. 2011-12-01T10:42:48 Kommander: Did you do something so that the diffusion coefficient for water is 0? Maybe I am doing that bit wrong and slowing mine down. 2011-12-01T10:43:52 gazpachoking: I don't diffuse if it is a water tile 2011-12-01T10:44:03 don't diffuse from it or to it 2011-12-01T10:44:18 Yeah, that's what I'm doing, just not sure if I am doing it right. 2011-12-01T10:44:35 I make another map of diffusion coefficients, with water having 0 2011-12-01T10:45:00 mine has water at 0, yes 2011-12-01T10:45:27 since I 0 out my map between each diffusion, I get water at 0 2011-12-01T10:47:45 rwest: Hmm, you zero out the heat of the water tiles between each diffusion? 2011-12-01T10:48:04 Doesn't that mean they are still absorbing heat? 2011-12-01T10:48:14 well I put my newly diffused values into a new map 2011-12-01T10:48:22 that starts as all 0 2011-12-01T10:48:50 since I add nothing to water tiles, diffusion does not go through them 2011-12-01T10:48:52 *** jon1 has quit IRC (Quit: Leaving.) 2011-12-01T10:49:38 what are the benefits from doing it that way? 2011-12-01T10:51:09 my map is a 2d array, so I put the values into a new array to prevent previous values calculated on that pass from effecting new ones 2011-12-01T10:52:38 *** alehorst has joined #aichallenge 2011-12-01T10:53:05 *** aerique has quit IRC (Quit: ...) 2011-12-01T10:53:07 my bot was favoring the bot right side of the map otherwise lol 2011-12-01T10:54:42 lol how? 2011-12-01T10:55:53 python is using 100% of quadcore... this can't be right 2011-12-01T10:57:52 *** alehorst has left #aichallenge 2011-12-01T10:59:37 http://ants.fluxid.pl/replay.7733 did my bots just do that? 2011-12-01T11:00:35 i mean my ants* lol :) 2011-12-01T11:01:09 *** george_ has joined #aichallenge 2011-12-01T11:01:30 does anyone know about the map symmetry types? 2011-12-01T11:01:48 i'm trying to understand them ... 2011-12-01T11:03:23 george_: are you familiar with git? 2011-12-01T11:04:23 yes 2011-12-01T11:05:04 ruuhkis: look what my bot did: http://ants.fluxid.pl/replay.7715?turn=180&row=80&col=72 2011-12-01T11:05:16 george_: go in the main repo 2011-12-01T11:05:20 george_: checkout amstan_mapgen 2011-12-01T11:05:28 *** jon1 has joined #aichallenge 2011-12-01T11:05:42 george_: run ./symmetricmap.py from ants/mapgen 2011-12-01T11:05:57 george_: feel free to open that file and play with the settings 2011-12-01T11:06:05 Minthos, I am jellybeans :( 2011-12-01T11:06:22 :) 2011-12-01T11:06:39 i am skeleton jelly 2011-12-01T11:06:55 a1k0n: no. you're a skeletoff 2011-12-01T11:07:07 then your bot all the sudden timed out 2011-12-01T11:07:10 but you still win :E 2011-12-01T11:07:32 I stopped it to run a new version 2011-12-01T11:07:37 I didn't know it was winning 2011-12-01T11:07:43 lol :D 2011-12-01T11:07:46 I feel you man(?) 2011-12-01T11:07:56 yes 2011-12-01T11:08:18 but your bot has tactics :( mines not smart enough 2011-12-01T11:08:21 *** george__ has joined #aichallenge 2011-12-01T11:08:29 ... 2011-12-01T11:08:46 sorry, seemed to mess up my browser 2011-12-01T11:09:03 *** george_ has quit IRC (Ping timeout: 265 seconds) 2011-12-01T11:09:10 what did i miss after 'are you familiar with git?' 2011-12-01T11:09:30 17:05 <@amstan> george_: go in the main repo 2011-12-01T11:09:30 17:05 <@amstan> george_: checkout amstan_mapgen 2011-12-01T11:09:30 17:05 -!- jon1 [~Adium@brick-lane.lbi.co.uk] has joined #aichallenge 2011-12-01T11:09:30 17:05 <@amstan> george_: run ./symmetricmap.py from ants/mapgen 2011-12-01T11:09:31 17:05 <@amstan> george_: feel free to open that file and play with the settings 2011-12-01T11:09:51 amstam. thanks. i guess i should do that then. 2011-12-01T11:10:03 @topic 2011-12-01T11:10:03 avdg: AI Challenge (sponsored by Google): http://aichallenge.org || Channel Logs: http://contestbot.aichallenge.org || Code Repo: http://github.com/aichallenge/aichallenge || About the game (Ants): http://aichallenge.org/problem_description.php || Submission deadline: December 18 23:59 EST 2011-12-01T11:10:12 is that code fixed as being the one in the final competition? 2011-12-01T11:10:15 ^check topic for gamelogs as well ;-) 2011-12-01T11:10:25 I mean channellogs 2011-12-01T11:10:39 * avdg gets confused again 2011-12-01T11:11:33 does the fluxid server have 5000ms of cpu time? 2011-12-01T11:11:46 yes but please don't use all of it 2011-12-01T11:11:53 turn time 2011-12-01T11:12:02 inc roundtrip packages 2011-12-01T11:12:06 it slows down games incredibly if bots take 5 seconds per turn 2011-12-01T11:12:08 (through network) 2011-12-01T11:12:11 i am usually 200-300 usually, but sometimes i still time out, which is odd 2011-12-01T11:12:17 using* 2011-12-01T11:12:20 it has 2:09.53 of cpu time if interested 2011-12-01T11:12:37 neat 2011-12-01T11:13:18 Minthos, thanks for relaying amstam's message 2011-12-01T11:13:30 np 2011-12-01T11:14:54 *** Ikol has joined #aichallenge 2011-12-01T11:15:43 *** Rav3nsW0rd has quit IRC (Ping timeout: 248 seconds) 2011-12-01T11:16:04 *** bretep has joined #aichallenge 2011-12-01T11:16:47 *** Ikol is now known as Rav3nSw0rd 2011-12-01T11:19:22 *** Rav3nSw0rd has quit IRC (Max SendQ exceeded) 2011-12-01T11:21:18 *** Rav3nSw0rd has joined #aichallenge 2011-12-01T11:27:51 *** Rav3nSw0rd has quit IRC (Quit: Leaving) 2011-12-01T11:28:13 *** Rav3nsW0rd has joined #aichallenge 2011-12-01T11:31:33 *** Rav3nsW0rd has quit IRC (Read error: Connection reset by peer) 2011-12-01T11:33:21 hkraal: amstan says I should talk with you :) 2011-12-01T11:33:56 janzert, no he said I should talk to you ;) 2011-12-01T11:33:59 oh, there's actually a C and C++ JIT based on clang called cling 2011-12-01T11:34:07 heh, there's a lot of neat stuff in llvm now :) 2011-12-01T11:35:54 well it's actually made by CERN/fermilab... 2011-12-01T11:36:24 http://root.cern.ch/drupal/content/what-cling 2011-12-01T11:37:24 http://ants.fluxid.pl/replay.7698 lol, 800 turns of doing nothing.... 2011-12-01T11:37:45 *** amstan has quit IRC (Ping timeout: 252 seconds) 2011-12-01T11:37:55 Fluxid: my hill died so my ants stopped gathering food and looking for it :) 2011-12-01T11:38:21 i have similar logic 2011-12-01T11:38:22 there was no hope of killing you though :( 2011-12-01T11:38:30 stupid 1 wide chokepoint 2011-12-01T11:38:47 I hate that map 2011-12-01T11:38:53 and the other random walks :P 2011-12-01T11:39:00 but ants still try to keep area visible, i must change their roles to attackers... that's a good thing to do next 2011-12-01T11:39:28 spec mentiones something about traversing with 3x3 square 2011-12-01T11:39:38 but "details are being worked on" 2011-12-01T11:40:03 my defense is basically broken 2011-12-01T11:40:21 your ants at least panic when enemy is near hive 2011-12-01T11:40:37 you can easily push my defenders away from hill and raze it 2011-12-01T11:40:47 ah 2011-12-01T11:40:50 yeah they try to defend 2011-12-01T11:40:56 but there are a dozen flaws with how it works :P 2011-12-01T11:41:30 like I pick ants closer to my hill as defenders first, so the ones closest to the enemy won't end up defenders, but won't be able to leave 2011-12-01T11:41:40 I weight enemy ants near my hill more as they approach it 2011-12-01T11:41:47 yeah, I do that 2011-12-01T11:41:55 I need to increase the weight 2011-12-01T11:42:14 ah, cool idea 2011-12-01T11:42:30 atm I just set the initial distance they are on the map to enemies as the distance they are from the hill 2011-12-01T11:42:35 so it's not weighted very much 2011-12-01T11:42:48 if that makes any sense :) 2011-12-01T11:43:52 my current AI still loses to my prior generation AI 2011-12-01T11:43:55 it's so annoying 2011-12-01T11:45:32 *** lbeew has joined #aichallenge 2011-12-01T11:45:54 *** mj41 has quit IRC (Ping timeout: 245 seconds) 2011-12-01T11:45:56 hi, quick question. how do i start the python3 started with wall time 2011-12-01T11:52:47 *** lavalamp has joined #aichallenge 2011-12-01T11:54:24 *** praveen_ has joined #aichallenge 2011-12-01T11:59:07 *** McLeopold has joined #aichallenge 2011-12-01T12:00:04 *** Cyndre has quit IRC (Ping timeout: 245 seconds) 2011-12-01T12:04:57 lbeew: what do you mean? 2011-12-01T12:05:39 lbeew: http://docs.python.org/py3k/library/time.html#time.time is this what you want? 2011-12-01T12:08:14 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-12-01T12:08:52 thats it, thanks 2011-12-01T12:23:47 *** olexs has joined #aichallenge 2011-12-01T12:24:04 In Game: Playing in a game right now. <- finally :) 2011-12-01T12:24:13 *** TheLinker has joined #aichallenge 2011-12-01T12:25:58 *** Blkt has quit IRC (Quit: ERC Version 5.3 (IRC client for Emacs)) 2011-12-01T12:29:06 janzert: call me naive, but shouldn't the sum of the gpm for each server total the gpm of the whole system? 2011-12-01T12:31:03 *** Surya has quit IRC (Quit: Page closed) 2011-12-01T12:36:41 It does? Doesn't it? 2011-12-01T12:37:19 3.^8^-1 + ... 2.7^-1 = 2.86games per minute 2011-12-01T12:37:25 *** Antimony has joined #aichallenge 2011-12-01T12:37:50 figures I break my battle code (completely removed its weight) and my bot does better 2011-12-01T12:38:10 by just ignoring enemy ants 2011-12-01T12:38:22 It's so cool to see my ants suddenly start ignoring food and converge on the enemy hill. http://aichallenge.org/visualizer.php?game=151197&user=3845 2011-12-01T12:38:53 *** Hexren has joined #aichallenge 2011-12-01T12:39:04 *** Hexren has left #aichallenge 2011-12-01T12:42:46 Antimony: it still could collect food along the way, it misses quite a lot of it... not really relevant in that one game, but in multi-hill games 2011-12-01T12:43:14 olexs: hi 2011-12-01T12:43:26 UncleVasya: hey 2011-12-01T12:43:46 Wanna talk about testing tool? 2011-12-01T12:43:51 yep, why not 2011-12-01T12:44:58 btw, I remember your bot was called Leviathan but I don't see it on the server. Have you renamed it or haven't uploaded yet? 2011-12-01T12:45:11 olexs, it avoids food to get to the hill faster 2011-12-01T12:45:24 the weighting between food and hills depends on current number of ants 2011-12-01T12:45:40 UncleVasya: it's olex.s now, and the current version is rubbish, apparently 2011-12-01T12:46:45 Antimony: ok, makes sense 2011-12-01T12:48:06 *** sorki is now known as rmarko_gone 2011-12-01T12:48:17 *** rmarko_gone is now known as sorki 2011-12-01T12:49:25 *** Antimony_ has joined #aichallenge 2011-12-01T12:50:15 So your question was why I don't upload it on forums. I need to do some work to make it finished (few bugfixes, few improvements). At one moment I realised I can not refuse any more the seduction to start writing the bot hard :) 2011-12-01T12:50:45 *** JorgeB has joined #aichallenge 2011-12-01T12:51:24 UncleVasya: did you think about maybe putting the source up on github? that way others could help you build it and implement new features faster, without you having to take your time for it 2011-12-01T12:52:13 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-12-01T12:52:15 It's written on Delphi. 2011-12-01T12:52:49 so? :) not the most popular language around, but it's not too hard to figure out 2011-12-01T12:53:20 *** liberforce has left #aichallenge 2011-12-01T12:53:59 *** kire has joined #aichallenge 2011-12-01T12:54:25 *** praveen_ has quit IRC (Ping timeout: 265 seconds) 2011-12-01T12:54:47 *** Antimony_ has quit IRC (Ping timeout: 258 seconds) 2011-12-01T12:55:17 Now I stuck in developing bot and need some rest so I have a time to do some work on AntTesting if still needed and not too late. 2011-12-01T12:56:13 too late it definitely isn't... and the tool will definitely be welcome by many on the forums, if it works with the current tools (playgame.py). 2011-12-01T12:56:19 *** Antimony has joined #aichallenge 2011-12-01T12:56:36 olexs: I will not implement everything that I wanted to do but I can make it work for you. 2011-12-01T12:57:08 As about git, I'll try to use it next time. 2011-12-01T12:57:39 that'd be great. honestly, I don't think it needs much more than it had before - a list of maps, a list of bots, a few setup options for playgame.py, and a table with game results 2011-12-01T12:57:50 *** kilae has joined #aichallenge 2011-12-01T12:58:38 *** elderwol1 is now known as Elderwolf 2011-12-01T12:58:57 I'll download tools.zip tomorrow (have no much traffic left for today). 2011-12-01T13:01:19 *** Antimony has quit IRC (Ping timeout: 248 seconds) 2011-12-01T13:02:02 olexs: InsaneMalkavian told me he gave you AntsTesting and it doesn't work. Was it version 0.5? 2011-12-01T13:02:44 UncleVasya: oh... not sure, I have the file on another pc, but I think not 2011-12-01T13:03:03 lol, copying my state object takes like 70ms 2011-12-01T13:03:09 note to self, don't do that ever 2011-12-01T13:03:10 *** ikaros has joined #aichallenge 2011-12-01T13:03:43 antimatroid: language? 2011-12-01T13:03:51 c++ 2011-12-01T13:04:18 I had a similar problem with java, although not quite so extreme :) 2011-12-01T13:04:30 olexs: Wanna some sparring bot for local testing? 2011-12-01T13:04:40 UncleVasya: yeah, why not 2011-12-01T13:07:21 AntsTesting 0.5 : http://www.filefactory.com/file/ce6905e/n/AntsTesting_0_5.7z 2011-12-01T13:07:32 Sorry, this file is no longer available. It may have been deleted by the uploader, or has expired. 2011-12-01T13:09:26 ok 2011-12-01T13:11:04 omg I actually one a maze match on tcp 2011-12-01T13:11:15 rewriting the diffusion code worked 2011-12-01T13:11:45 rwest, which bot ? 2011-12-01T13:11:55 http://tcpants.com/replay.14116 2011-12-01T13:12:44 olexs: This is the current version on the site: http://filefactory.com/file/cf64751/n/UncleVasya_22_1_Olga.7z 2011-12-01T13:12:51 Place 18. 2011-12-01T13:12:59 http://aichallenge.org/profile.php?user=53 2011-12-01T13:13:00 thanks. 2011-12-01T13:13:01 *** bhasker has quit IRC (Ping timeout: 240 seconds) 2011-12-01T13:13:48 Let me know when you'll overpower her. 2011-12-01T13:14:29 will do. have to run now again, sadly... see you later 2011-12-01T13:14:48 Tomorrow I think. gl 2011-12-01T13:16:12 UncleVasya: boooo I was gonna try beating that bot 2011-12-01T13:16:14 *** jon1 has quit IRC (Quit: Leaving.) 2011-12-01T13:16:16 but its a .exe 2011-12-01T13:16:17 lol 2011-12-01T13:16:44 wine? 2011-12-01T13:16:57 g0llum: bout to try that 2011-12-01T13:17:52 honestly, i didn't finf the link on that crapsite 2011-12-01T13:18:00 *** Harpyon has joined #aichallenge 2011-12-01T13:18:09 wine isn't in debian repos... really lol 2011-12-01T13:19:36 not even available pre-built 2011-12-01T13:19:45 *** amstan has joined #aichallenge 2011-12-01T13:19:45 *** ChanServ sets mode: +o amstan 2011-12-01T13:20:01 g0llum: When I ask to tell me a good place for upload I get dropbox as an answer. Due to some reasons I need place that works like that: 2011-12-01T13:20:21 box.net 2011-12-01T13:20:24 ok, i see. 2011-12-01T13:20:39 open page in browser --> click button Browse and pick file --> click button Upload ---> get link 2011-12-01T13:20:47 if you just need to do a quick upload you can't beat ompldr 2011-12-01T13:20:57 and there's even a nice CLI client you can pipe stuff into 2011-12-01T13:20:59 http://ompldr.org/ 2011-12-01T13:21:17 just post to the forum & attach a zip 2011-12-01T13:21:32 UncleVasya: dropbox: put file in folder, right click=> get shareable link 2011-12-01T13:22:46 *** Zannick has quit IRC (Ping timeout: 258 seconds) 2011-12-01T13:23:22 UncleVasya, sorry for being harsh, didn't want that. 2011-12-01T13:23:24 I've heard dropbox is greedy to traffic. My reason is not my stupidness, but the limit of 20 mb per day. 2011-12-01T13:24:10 I don't feel like you must be sorry for something so never mind. 2011-12-01T13:24:37 20 mb per day? :0 2011-12-01T13:24:48 dropbox will probably burn through that, I agree 2011-12-01T13:24:49 I've never experienced this 20mb limit 2011-12-01T13:24:52 http://aichallenge.org/visualizer.php?game=151355&user=3473 2011-12-01T13:25:06 that version may have bugs but it did snatch victory from the jaws of defeat 2011-12-01T13:25:08 But it's not the speediest service 2011-12-01T13:25:16 *** grwip has quit IRC (Remote host closed the connection) 2011-12-01T13:25:26 contestbot: seen janzert 2011-12-01T13:25:26 amstan: janzert was last seen in #aichallenge 1 hour, 52 minutes, and 4 seconds ago: hkraal: amstan says I should talk with you :) 2011-12-01T13:25:45 UncleVasya: oh, I see what you meant about the 20 mb 2011-12-01T13:26:59 janzert, hkraal: so what's up? 2011-12-01T13:27:37 need to go, bye. 2011-12-01T13:27:40 *** UncleVasya has quit IRC () 2011-12-01T13:29:32 *** Zannick has joined #aichallenge 2011-12-01T13:33:27 so I think I just realized that my bot needs a concept of stalemates 2011-12-01T13:33:28 *** Antimony has joined #aichallenge 2011-12-01T13:33:43 from my earliest work I always kind of assumed you could just throw more ants at a problem and it would go away 2011-12-01T13:34:02 BenJackson: how do you evaluate whether attacking would be a good idea? 2011-12-01T13:34:11 I get a int with a score for a given situation 2011-12-01T13:34:40 yeah, I score based on how many ants I have and how well defended it is 2011-12-01T13:34:44 plus misc special cases 2011-12-01T13:35:10 it finds it very important to kill off very close hills, too 2011-12-01T13:35:16 so if you control most of the map and you have a bunch of ants, will you sacrifice more ants in a battle? 2011-12-01T13:35:29 but if it will stalemate in a 3-wide hallway then I should just seal it off and go on with life 2011-12-01T13:35:38 I guess it's a good idea to build up for a bit if you control lots of territory 2011-12-01T13:35:54 I will sacrifice with more ants, but I haven't really quantified what that rule does 2011-12-01T13:36:04 *** choas has joined #aichallenge 2011-12-01T13:36:09 yeah, I haven't decided on a good way to adjust how many I'll sacrifice 2011-12-01T13:36:58 I just have some dumb hardcoded rules atm 2011-12-01T13:37:20 http://aichallenge.org/visualizer.php?game=150587&user=3473&turn=133&row=2&col=114 2011-12-01T13:37:32 if I had just accepted that stalemate there and ignored it I would have been fine 2011-12-01T13:38:01 but my game logic dates back to pre-combat suicide marches so it just figures if it throws more ants in the clog will clear up 2011-12-01T13:38:27 ah 2011-12-01T13:38:27 BenJackson: you must have some complicated code for fighting 2011-12-01T13:38:42 BenJackson: and you should put your bot with attack code back on bhickey.net 2011-12-01T13:38:58 your ants keep running from me 2011-12-01T13:39:00 rwest: I bet his fighting is actually really simple :) 2011-12-01T13:39:04 simple things usually work better 2011-12-01T13:39:43 thestinger: mine uses a bfs to find adjacent ants and weight them correctly, it's evil 2011-12-01T13:39:53 I didn't wanna use one 2011-12-01T13:40:47 I cannot win a single damn game on fluxid 2011-12-01T13:42:03 I reuploaded to the site yesterday 2011-12-01T13:42:19 http://aichallenge.org/profile.php?user=2255 climbing back up to where I was (and hopefully higher :P) 2011-12-01T13:42:37 *** p_l has quit IRC (Ping timeout: 240 seconds) 2011-12-01T13:43:04 my ancient version was ~81-82 skill and hovered around rank 20 to rank 50 2011-12-01T13:43:21 the fluxid bots are so much better than the top 100 on the official site :P 2011-12-01T13:43:35 my old version couldn't win any games on fluxid anymore 2011-12-01T13:44:03 *** Redgis has joined #aichallenge 2011-12-01T13:44:12 you say it like i made all the bots ;P 2011-12-01T13:44:33 it's just the easy way to refer to the url :) 2011-12-01T13:44:42 ;) 2011-12-01T13:44:53 18 more days 2011-12-01T13:45:08 i must reupload it at least one week before 18th... 2011-12-01T13:45:12 so little time 2011-12-01T13:46:15 yep 2011-12-01T13:53:27 woah, I just looked at my code and noticed I was attacking ants only when I was outnumbered, not the other way around 2011-12-01T13:53:30 haha 2011-12-01T13:55:28 introducing awesome bugs ftl 2011-12-01T13:55:29 how do you guys prevent timeouts, check time_remaining regularly? 2011-12-01T13:56:00 yep, pretty much 2011-12-01T13:56:04 lbeew: I don't check time at all atm 2011-12-01T13:56:04 lbeew: I don't come close enough to timing out to check, so I dunno 2011-12-01T13:56:21 *** dvladim has joined #aichallenge 2011-12-01T13:56:32 do you guys use A* search algorithms? 2011-12-01T13:56:34 I just rewrote my combat in a way where it can't timeout, and I'll add combat that uses up the rest of the turn afterwards 2011-12-01T13:56:46 My bot hasn't needed it lately either, but it's nice to have in case something weird happens 2011-12-01T13:56:50 my current bots only use ~20-40ms with 100+ ants in battles 2011-12-01T13:56:57 wow 2011-12-01T13:57:10 what language are you coding in 2011-12-01T13:57:11 pre-combat is 4ms to 10ms 2011-12-01T13:57:13 C++ 2011-12-01T13:57:25 what kind of path finding algorithm? 2011-12-01T13:57:48 sort of a BFS to make distance maps, but with all kinds of hacks 2011-12-01T13:58:16 I guess it's like the reverse of the collaborative diffusion stuff 2011-12-01T13:58:34 incrementing values away from targets instead of spreading/diffusing them 2011-12-01T13:59:18 with just a BFS to get a distance map you can get distances/shortest paths for all tiles, but that doesn't take into account collaborative movement (it assumes ants don't block each other) 2011-12-01T14:00:07 still nice 2011-12-01T14:00:40 yea the mleitung factor 2011-12-01T14:01:10 i've implemented an A* algorithm, but it's so time consuming, also not sure how to save the paths 2011-12-01T14:01:27 A* for each ant to each target was way to slow for me 2011-12-01T14:02:04 yeah I don't use a* either, just bfs 2011-12-01T14:02:11 even in C++ and somewhat optimized it was too slow for 300+ ants 2011-12-01T14:02:13 yes, it's very expensive so if i get more than 20 some ants it times out 2011-12-01T14:02:19 and it used way too much of my turn 2011-12-01T14:02:26 i find it hard to squeese 40 ants through a gap 3 tiles wide.. any i deas on that ? 2011-12-01T14:03:03 g0llum: what will your ants do? 2011-12-01T14:03:18 good question 2011-12-01T14:03:42 mine will fill the whole pathway, and they can swap places with each other 2011-12-01T14:03:54 so they'll get through it in a fairly efficient manner 2011-12-01T14:04:06 guess, if isolve that, former question is solved 2011-12-01T14:04:08 *** kire has quit IRC (Remote host closed the connection) 2011-12-01T14:09:21 *** amstan has quit IRC (Remote host closed the connection) 2011-12-01T14:11:17 *** amstan has joined #aichallenge 2011-12-01T14:11:17 *** ChanServ sets mode: +o amstan 2011-12-01T14:12:15 I have too many branches :( 2011-12-01T14:12:21 I have no idea what's going on anymore 2011-12-01T14:13:22 *** thestinger has quit IRC (Quit: WeeChat 0.3.6) 2011-12-01T14:13:27 merge them all at once! what could possibly go wrong 2011-12-01T14:13:34 or... never mind 2011-12-01T14:13:44 =) 2011-12-01T14:14:39 *** ruuhkis has quit IRC () 2011-12-01T14:15:18 *** thestinger has joined #aichallenge 2011-12-01T14:15:28 thestinger: merge them all at once! what could possibly go wrong 2011-12-01T14:15:39 :) 2011-12-01T14:16:22 I like testing them on the tcp server first but it's too slow 2011-12-01T14:18:31 *** lawmax has quit IRC (Ping timeout: 265 seconds) 2011-12-01T14:19:00 *** lbeew has quit IRC (Ping timeout: 265 seconds) 2011-12-01T14:19:00 *** george__ has quit IRC (Ping timeout: 265 seconds) 2011-12-01T14:19:29 *** bretep has quit IRC (Ping timeout: 265 seconds) 2011-12-01T14:19:58 *** lavalamp has quit IRC (Ping timeout: 265 seconds) 2011-12-01T14:22:05 luck for all, noone seriously tried genetic algo's there 2011-12-01T14:25:36 I just cleaned up some old unused code, reduced my main source file to 500 lines 2011-12-01T14:25:52 but discovered my combat file was 600 lines alone :/ 2011-12-01T14:26:12 got it down to 500, but it's still bigger than I thought 2011-12-01T14:27:30 rank/sizeof(zipped_code) would be a really sportive thing 2011-12-01T14:28:39 *** sofuture is now known as modafinil 2011-12-01T14:28:46 *** olexs has left #aichallenge 2011-12-01T14:40:12 *** dvladim has quit IRC (Ping timeout: 258 seconds) 2011-12-01T14:41:03 srry for sounding grumpy, it's all for not getting hired by some game company here, where the *lead* came up with ~8000 lines in main.cpp 2011-12-01T14:42:11 :) 2011-12-01T14:43:42 wouldn't you rather work somewhere with people you respect and admire? :) 2011-12-01T14:43:45 missed your chance at slaving away 80h/week for a middle-end wage 2011-12-01T14:44:13 licking my wounds, and saying, for the better.. 2011-12-01T14:45:56 http://aichallenge.org/visualizer.php?game=151363&user=757 two interesting things in this game 2011-12-01T14:46:12 one is that i don't gather food in cave just right by my hive 2011-12-01T14:46:22 i don't keep area visible at all, wtf 2011-12-01T14:46:40 and second, right before flag timeout, you can see him stutter 2011-12-01T14:48:28 maybe his combat code is too expensive for official server 2011-12-01T14:48:52 still, i saw him manage with at least 300 ants 2011-12-01T14:49:01 but here he has got only 164 2011-12-01T14:49:22 he even pauses for few turns 2011-12-01T14:49:44 or move only few of his ants 2011-12-01T14:50:07 pause could be a sign of starvation, his bot detects timeout is imminent and aborts everything 2011-12-01T14:50:29 i know that, i just wonder what could go wrong if not combat 2011-12-01T14:50:49 he doesn't even have many enemy ants to deal with 2011-12-01T14:51:29 anyway i like this map. looks like intestine, and has those funny caves 2011-12-01T14:52:05 yeah, my bot is completely broken on server 2011-12-01T14:52:15 i wonder how it managed to up to 16th place 2011-12-01T14:58:24 *** Antimony has quit IRC (Ping timeout: 245 seconds) 2011-12-01T14:59:00 *** Palmik has quit IRC (Remote host closed the connection) 2011-12-01T15:02:16 *** Antimony has joined #aichallenge 2011-12-01T15:02:51 *** QuirionPT has joined #aichallenge 2011-12-01T15:04:51 *** Accoun has quit IRC () 2011-12-01T15:17:39 *** Blkt has joined #aichallenge 2011-12-01T15:22:07 *** Redgis has quit IRC (Read error: Connection reset by peer) 2011-12-01T15:22:54 *** Redgis has joined #aichallenge 2011-12-01T15:25:34 *** epicmonkey has joined #aichallenge 2011-12-01T15:29:23 *** oxttp has joined #aichallenge 2011-12-01T15:29:28 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-12-01T15:30:21 *** QuirionPT has quit IRC (Quit: Leaving) 2011-12-01T15:31:24 *** oxttp has left #aichallenge 2011-12-01T15:39:01 *** Antimony has quit IRC (Ping timeout: 252 seconds) 2011-12-01T15:45:24 *** Accoun has joined #aichallenge 2011-12-01T15:45:54 *** JorgeB has quit IRC (Quit: Computer has gone to sleep.) 2011-12-01T15:50:51 Fluxid: my version 7 wandered from 16-22nd place and my new version is clearly better 2011-12-01T15:51:02 but I don't know that the game rate on the server would have ever pushed the old one down if I didn't resubmit 2011-12-01T15:51:18 I think they algorithm they're using just won't converge for such small game rates 2011-12-01T15:51:49 I have no idea how they're going to make the "finals" work 2011-12-01T15:52:01 they don't have nearly enough compute horsepower 2011-12-01T15:52:09 my bot is prepared if turntime gets cut down to 50ms :) 2011-12-01T15:52:40 switching to CPU time instead of wall time would help a lot 2011-12-01T15:52:58 you could run multiple matches on one instance and soak up all the CPU 2011-12-01T15:54:01 50ms would be brutal 2011-12-01T15:54:15 50ms would be a whole new contest 2011-12-01T15:54:20 seriously 2011-12-01T15:54:43 I def wouldn't be using java haha 2011-12-01T15:55:44 50 ms locally for me isn't the same as 50 ms on the server 2011-12-01T15:56:37 *** Aloren has joined #aichallenge 2011-12-01T15:56:47 *** delt0r_ has quit IRC (Ping timeout: 248 seconds) 2011-12-01T15:57:08 *** Aloren has left #aichallenge 2011-12-01T16:06:45 *** kilae has quit IRC (Ping timeout: 260 seconds) 2011-12-01T16:10:02 *** delt0r_ has joined #aichallenge 2011-12-01T16:13:12 BenJackson: http://tcpants.com/replay.14283 you wanted pink's hill so bad 2011-12-01T16:13:56 *** amstan_ has joined #aichallenge 2011-12-01T16:13:57 *** ChanServ sets mode: +o amstan_ 2011-12-01T16:16:13 *** kilae has joined #aichallenge 2011-12-01T16:18:13 *** ltriant has joined #aichallenge 2011-12-01T16:20:19 *** jstrong has joined #aichallenge 2011-12-01T16:26:34 *** Antimony has joined #aichallenge 2011-12-01T16:27:30 *** mikewintermute has joined #aichallenge 2011-12-01T16:32:48 *** AntDroid has joined #aichallenge 2011-12-01T16:38:09 *** mj41 has joined #aichallenge 2011-12-01T16:43:12 *** mleise has quit IRC (Ping timeout: 259 seconds) 2011-12-01T16:43:42 *** mikewintermute has quit IRC (Quit: mikewintermute) 2011-12-01T16:49:24 rwest: not enough! 2011-12-01T16:49:30 haha 2011-12-01T16:50:31 did something similarly suicidal on aichallenge too 2011-12-01T16:50:45 I'm saving games like that using the extra args: http://tcpants.com/replay.14283?turn=696&col=25&row=60 2011-12-01T16:50:50 putting them in my todo list 2011-12-01T16:51:45 I need to implement standing still as an valid move hah 2011-12-01T16:51:55 I think that may help a lot 2011-12-01T16:52:30 I see there are times when I am completely blocked and I just step on my own ants 2011-12-01T16:52:59 rare, but if I can't move in any direction and there is an ant north, squish 2011-12-01T16:55:53 I extended everything to consider move=4 to be still 2011-12-01T16:56:08 for a while I was even ordering ants to go "x" to the server, oops 2011-12-01T16:58:09 I can just move my movement into my map class and just not move if all possible moves are -100000 weighted (my score given to occupied, water or soon to be occupied tiles) 2011-12-01T16:58:32 I have a feeling there are a lot of times that not moving is my best move 2011-12-01T16:59:18 *** kilae has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111104165243]) 2011-12-01T16:59:23 *** JorgeB has joined #aichallenge 2011-12-01T17:01:10 *** Jak_o_Shadows has joined #aichallenge 2011-12-01T17:01:48 See if you can guess when my bot spotted the enemy hill. http://aichallenge.org/visualizer.php?game=151197&user=3845 2011-12-01T17:02:17 *** choas has quit IRC (Quit: leaving) 2011-12-01T17:03:49 *** AntDroid has quit IRC (Ping timeout: 265 seconds) 2011-12-01T17:06:39 *** amstan_ has quit IRC (Ping timeout: 248 seconds) 2011-12-01T17:08:36 the most fair 5 player map ever http://paste.aichallenge.org/dV8f8/ 2011-12-01T17:10:54 *** JorgeB has quit IRC (Quit: Computer has gone to sleep.) 2011-12-01T17:11:26 *** McLeopold has quit IRC (Quit: Leaving.) 2011-12-01T17:17:00 lol, random walks 2011-12-01T17:17:03 http://aichallenge.org/visualizer.php?game=151973&user=2255 2011-12-01T17:17:57 I die before getting any points and finish second :P 2011-12-01T17:18:13 well I mean my _hill_ dies 2011-12-01T17:19:09 McLeopold: Based on game analysis or some static quality of that map? 2011-12-01T17:19:35 I assume the latter 2011-12-01T17:22:31 BenJackson: I guess because all starting positions are identical 2011-12-01T17:27:44 haha, fantastic 2011-12-01T17:40:15 argh, in my one aichallenge game today one of my neighbors suicided into my hill 2011-12-01T17:40:22 this is sooo informative 2011-12-01T17:42:47 I wonder if that's a viable strategy overall 2011-12-01T17:43:01 only goal being to trade your 1 point hills for 2 point hill caps 2011-12-01T17:43:03 random walks are teh useless 2011-12-01T17:43:10 *** jstrong is now known as roflmao 2011-12-01T17:43:17 BenJackson: suiciding into hills? that's a perfectly valid tactic :) 2011-12-01T17:43:39 I should run my old bot without combat on tcp again 2011-12-01T17:43:50 thestinger: in this case by "suicide" I mean he abandoned his hill and sent all his ants to mine 2011-12-01T17:44:01 ah, I still do that 2011-12-01T17:44:01 lost his hill about the same time he took mine 2011-12-01T17:44:04 but I don't stop foraging 2011-12-01T17:44:44 meh, this territory control stuff isn't working at all 2011-12-01T17:44:50 I try to learn things from these games, but so many things are outside your control 2011-12-01T17:45:00 one ant happens to only go one way so his neighbor expands twice as fast 2011-12-01T17:48:30 *** NoxiaZ^ has quit IRC (Read error: Connection reset by peer) 2011-12-01T17:49:04 *** Blkt has quit IRC (Remote host closed the connection) 2011-12-01T17:56:57 *** McLeopold has joined #aichallenge 2011-12-01T18:03:11 *** Antimony has quit IRC (Ping timeout: 244 seconds) 2011-12-01T18:07:47 anyone have a good conecpt for group engagements? 2011-12-01T18:12:23 *** Antimony has joined #aichallenge 2011-12-01T18:15:20 a1k0n: http://www.reddit.com/r/programming/comments/mwk5u/7_reasons_that_rexx_still_matters/c34g9zf 2011-12-01T18:17:33 *** epicmonkey has quit IRC (Ping timeout: 258 seconds) 2011-12-01T18:18:26 haha 2011-12-01T18:23:44 jab_bott's visualizer is really nice 2011-12-01T18:23:48 just downloaded it 2011-12-01T18:23:50 http://ompldr.org/vYmpveg 2011-12-01T18:24:28 thestinger: where do I get that? 2011-12-01T18:24:52 rwest_: https://github.com/j-h-a/aichallenge/blob/vis_overlay/VIS_OVERLAY.md 2011-12-01T18:25:31 it's quite handy, but beware the "info" to put in text 2011-12-01T18:25:42 I think whoever added that part did something wrong 2011-12-01T18:25:46 becasue if you use it it is HORRIBLY slow 2011-12-01T18:26:41 Oh yeah, that's nice. I gotta check that out. 2011-12-01T18:30:42 *** TheLinker has quit IRC (Remote host closed the connection) 2011-12-01T18:31:57 *** yoden has joined #aichallenge 2011-12-01T18:32:54 *** cyphase_ has joined #aichallenge 2011-12-01T18:32:58 lol it made me timeout 2011-12-01T18:33:03 gotta increase that 2011-12-01T18:33:12 I have to set turntime really high to do lots of drawing 2011-12-01T18:33:25 *** cyphase has quit IRC (Ping timeout: 240 seconds) 2011-12-01T18:33:33 omg slow 2011-12-01T18:33:33 lol 2011-12-01T18:33:51 did the tcp server just reset? 2011-12-01T18:34:08 tcpants seems down 2011-12-01T18:34:16 if thats the one you are talking aboutr 2011-12-01T18:34:43 that's the one 2011-12-01T18:39:27 rwest_: you might have to hack engine.py to allow more lines/sec 2011-12-01T18:39:36 it has a builtin 10ms sleep between every 100 lines of bot output 2011-12-01T18:39:42 with the visualizer that can waste a ton of time 2011-12-01T18:40:25 *** amstan_ has joined #aichallenge 2011-12-01T18:40:38 *** ChanServ sets mode: +o amstan_ 2011-12-01T18:42:59 I can't get the colors to work 2011-12-01T18:43:09 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-12-01T18:44:11 you have to set the fill color or whatever 2011-12-01T18:44:17 and then fill tiles with it, etc. 2011-12-01T18:44:51 rofl I probable should fill the tiles after setting the color 2011-12-01T18:44:53 hahahaha 2011-12-01T18:45:29 *** mj41 has quit IRC (Ping timeout: 255 seconds) 2011-12-01T18:46:29 if you want to paitn the whole map you should set the color, emit "tile" for everything that color, repeat 2011-12-01T18:46:36 not color,tile,color,tile 2011-12-01T18:46:42 it's slow enough with just tiles... 2011-12-01T18:46:54 where is engine.py? 2011-12-01T18:47:35 got it 2011-12-01T18:48:59 *** HaraKiri has quit IRC () 2011-12-01T18:50:23 BenJackson: lol, I'm doing it with color,tile,color,tile etc. 2011-12-01T18:50:26 fast enough :) 2011-12-01T18:50:34 end up with a 50MB html file though 2011-12-01T18:50:43 I am too, mine is slow as shit though 2011-12-01T18:50:48 what CPU do you have? 2011-12-01T18:50:58 phenom 2 x4 2011-12-01T18:51:46 almost 2s per turn lol 2011-12-01T18:56:01 *** ikaros has joined #aichallenge 2011-12-01T18:56:38 and just 200 turns is taking forever to load the file 2011-12-01T18:57:26 *** Antimony has quit IRC (Ping timeout: 244 seconds) 2011-12-01T18:58:12 *** amstan_ has quit IRC (Read error: No route to host) 2011-12-01T19:01:01 rwest_: the engine has a 10ms delay per 50 lines or something like that by default, you can turn that off 2011-12-01T19:01:23 I did 2011-12-01T19:02:34 run the engine with pypy? :P 2011-12-01T19:03:00 I think I may take a shower while running a 1k turns game 2011-12-01T19:03:19 you'll end up with a 3GB json file... 2011-12-01T19:03:30 haha 2011-12-01T19:03:52 *** cyphase_ is now known as cyphase 2011-12-01T19:04:06 *** cyphase has joined #aichallenge 2011-12-01T19:04:30 *** bhasker has joined #aichallenge 2011-12-01T19:10:48 *** besh has joined #aichallenge 2011-12-01T19:11:14 so 2011-12-01T19:11:16 in order words 2011-12-01T19:11:21 lets all put our big brains together 2011-12-01T19:11:24 and optimize the visualizer code :P 2011-12-01T19:11:43 I'm pretty sure using alphas is what brings it to its knees 2011-12-01T19:11:50 like 75.4% sure 2011-12-01T19:12:28 *** besh has quit IRC (Client Quit) 2011-12-01T19:13:08 *** Harpyon has quit IRC (Quit: Textual IRC Client: http://www.textualapp.com/) 2011-12-01T19:13:55 is the actual visualizer slow for you guys? 2011-12-01T19:14:28 or just running the game 2011-12-01T19:14:44 running the game 2011-12-01T19:14:50 and loading the visualizer 2011-12-01T19:20:47 also if youre getting huge file sizes 2011-12-01T19:20:51 js is going to be slwo to load that 2011-12-01T19:20:54 i bet it can be reduced 2011-12-01T19:21:15 *** Cyndre has joined #aichallenge 2011-12-01T19:23:01 1000 turns with it coming up haha 2011-12-01T19:24:22 i might actually look into the code ;P 2011-12-01T19:26:34 ok, who is the combatant 2011-12-01T19:27:00 that is a clever pun i must admit 2011-12-01T19:27:10 what are you talking about? 2011-12-01T19:27:17 http://ants.fluxid.pl/player/CombatAnt 2011-12-01T19:27:22 *** dom7b5 has quit IRC (Quit: Ex-Chat) 2011-12-01T19:27:43 *** dom7b5 has joined #aichallenge 2011-12-01T19:27:57 dang those are some skills 2011-12-01T19:28:26 huh new cell maze maps? 2011-12-01T19:28:55 I believe that would be amstan 2011-12-01T19:29:45 he was doing that pattern 2011-12-01T19:29:59 *** raemde_ has joined #aichallenge 2011-12-01T19:30:03 oh no way 2011-12-01T19:30:16 i thought it was mcleopold 2011-12-01T19:30:53 fluxid timing out 2011-12-01T19:30:54 oh nooo 2011-12-01T19:31:45 k i give up 2011-12-01T19:31:51 my bot only gets worse with every change i make 2011-12-01T19:32:08 i added min max logic and it sucks at combat now because my move generator is horrible:( 2011-12-01T19:33:01 *** raemde has quit IRC (Ping timeout: 240 seconds) 2011-12-01T19:33:11 *** kurti has quit IRC (Ping timeout: 255 seconds) 2011-12-01T19:33:19 *** sorki has quit IRC (Ping timeout: 252 seconds) 2011-12-01T19:33:37 *** sorki has joined #aichallenge 2011-12-01T19:33:46 *** Redgis has quit IRC (Ping timeout: 244 seconds) 2011-12-01T19:34:49 *** kurti has joined #aichallenge 2011-12-01T19:35:59 *** Antimony has joined #aichallenge 2011-12-01T19:48:44 *** Antimony has quit IRC (Ping timeout: 252 seconds) 2011-12-01T19:49:06 *** Garf has quit IRC (Quit: Make a new plan, Stan!) 2011-12-01T19:51:10 oh wow 2011-12-01T19:51:14 my foraging my totally broken 2011-12-01T19:51:19 * thestinger sighs 2011-12-01T19:51:29 my foraging was* totally broken 2011-12-01T19:52:38 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-12-01T19:53:33 *** TheLinker has joined #aichallenge 2011-12-01T19:53:55 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-12-01T19:57:32 *** avdg has quit IRC (Ping timeout: 244 seconds) 2011-12-01T19:59:39 *** analyst74 has joined #aichallenge 2011-12-01T20:01:26 I'm running all the versions of my bot against holdbots on every map 2011-12-01T20:01:30 to get some kind of exploration baseline 2011-12-01T20:01:34 so I can fix it once and for all 2011-12-01T20:01:39 what is a holdbot? 2011-12-01T20:01:46 just stays at start? 2011-12-01T20:01:47 one of the samples 2011-12-01T20:01:52 yeah, it doesn't go anywhere 2011-12-01T20:01:56 ah k 2011-12-01T20:02:23 my min max code for combat is completely fubared 2011-12-01T20:02:35 *** Antimony has joined #aichallenge 2011-12-01T20:02:42 the eval function needs a lot of work and stupid move generator blows 2011-12-01T20:03:02 *** b0rder has joined #aichallenge 2011-12-01T20:06:44 move generator could be slow but how could it be *bad*? 2011-12-01T20:06:50 unless you are only evaluating a subset of moves 2011-12-01T20:07:35 BenJackson: mine only evaluates 4 sets of moves per battle atm 2011-12-01T20:08:24 I can do more but I was worried about it taking too long and not being able to handle all the battles 2011-12-01T20:13:10 *** avdg has joined #aichallenge 2011-12-01T20:22:01 and it has begun, I am making scenarios 2011-12-01T20:22:47 contestbot: seen janzert 2011-12-01T20:22:47 amstan: janzert was last seen in #aichallenge 8 hours, 49 minutes, and 25 seconds ago: hkraal: amstan says I should talk with you :) 2011-12-01T20:22:56 janzert, hkraal: ping 2011-12-01T20:24:41 *** ThiagoRRamos has joined #aichallenge 2011-12-01T20:25:44 *** ThiagoRRamos has quit IRC (Changing host) 2011-12-01T20:25:44 *** ThiagoRRamos has joined #aichallenge 2011-12-01T20:28:07 *** AntDroid has joined #aichallenge 2011-12-01T20:34:48 Hey 2011-12-01T20:41:53 *** avdg has quit IRC (Ping timeout: 252 seconds) 2011-12-01T20:52:22 *** avdg has joined #aichallenge 2011-12-01T20:54:15 ThiagoRRamos: hi 2011-12-01T21:17:06 amstan: we talked for a bit and both had to take off, going to write him an email later to see about specifics and if we can make it work 2011-12-01T21:18:52 *** Cyndre has quit IRC (Ping timeout: 258 seconds) 2011-12-01T21:27:01 If my bot ends up writing more than 10 MB of data on the servers, how will I know? 2011-12-01T21:30:01 The bot will just timeout? 2011-12-01T21:33:49 *** Jak_o_Shadows has quit IRC (Remote host closed the connection) 2011-12-01T21:35:10 *** dai-ra_ has quit IRC (Read error: Operation timed out) 2011-12-01T21:35:15 *** dai-ra has joined #aichallenge 2011-12-01T21:35:26 janzert: what's the deal with hkraal? 2011-12-01T21:35:56 some possible server use donation 2011-12-01T21:36:34 good enough to replace all of the ec2 instances? 2011-12-01T21:36:34 ThiagoRRamos you almost certainly want to turn off almost all of your logging before you upload to the server. 2011-12-01T21:36:40 *** bmh has joined #aichallenge 2011-12-01T21:37:08 McLeopold: I'm hoping we could use both, but possibly 2011-12-01T21:37:21 *** antimatroid has quit IRC (Ping timeout: 244 seconds) 2011-12-01T21:37:29 I was wondering what your thoughts on that are 2011-12-01T21:38:19 all the bots in each game would be running on the same hardware obviously but some games would be getting better hardware than others 2011-12-01T21:38:39 bugnuts: Actually I can't use the log :P 2011-12-01T21:38:42 of course ec2 doesn't guarantee that all instances are exactly the same either 2011-12-01T21:39:11 bugnuts:This 10MB data limit just refers to the loggings? 2011-12-01T21:39:16 idk... 2011-12-01T21:39:38 McLeopold: also it's still pretty tentative on whether we can get it worked out to use at all 2011-12-01T21:40:56 *** Conorach has quit IRC (Ping timeout: 252 seconds) 2011-12-01T21:40:59 I think I'm okay with it before finals 2011-12-01T21:41:10 not necessarily logging but writing anything is a mistake I think 2011-12-01T21:41:39 beyond the required commands to stout that is 2011-12-01T21:41:55 hmm 2011-12-01T21:42:18 so this limit is actually the stdout? 2011-12-01T21:42:36 there is no limit on stdout. 2011-12-01T21:42:53 ThiagoRRamos: the 10mb limit is on writing files, not stdout or stderr 2011-12-01T21:43:09 There should be no good reason to create files 2011-12-01T21:43:23 Hmm 2011-12-01T21:43:32 I think thats right. it's a limit on things which turn up in the filesystem 2011-12-01T21:43:49 Hmm 2011-12-01T21:43:55 Hmm? 2011-12-01T21:43:58 I think I understood now 2011-12-01T21:44:07 yes, the 10MB limit is meant for things that actually end up on the disk 2011-12-01T21:44:08 It's that my bot timed out recently on a server game 2011-12-01T21:44:27 And this time I really couldn't find out a good time reason 2011-12-01T21:45:25 So I thought there sould be something wrong about this limit 2011-12-01T21:45:38 but actually I couldn't use a log yet... 2011-12-01T21:46:00 I find it amusing how a signle ant of mine just ate a complete trail of food all around the center of the map. http://aichallenge.org/visualizer.php?game=152688&user=3845&turn=219 2011-12-01T21:46:08 unless you need multiple files, I'd just write to stderr 2011-12-01T21:46:23 *** bhasker has quit IRC (Ping timeout: 255 seconds) 2011-12-01T21:47:23 Yeah, in Python I couldn't do it yet 2011-12-01T21:47:39 I could write to stdout 2011-12-01T21:47:54 so the engine would say there was an error en would print my message 2011-12-01T21:48:03 but it isn't very good 2011-12-01T22:01:23 janzert: he was talking about using virtualization to make his cpus dumber so we can match ec2 2011-12-01T22:01:29 janzert: won't that work? 2011-12-01T22:04:44 I don't have a problem with them being faster 2011-12-01T22:05:15 *** Antimony has quit IRC (Ping timeout: 244 seconds) 2011-12-01T22:11:06 janzert, amstan, McLeopold: are you guys getting a 501(c)3 set up? 2011-12-01T22:11:34 bmh: jeff was supposed to do that, but i haven't heard of any updates from him 2011-12-01T22:11:45 *facepalm* 2011-12-01T22:12:02 also, we wouldn't have anyone to join it 2011-12-01T22:12:10 he wanted to have it as a canadian non-profit 2011-12-01T22:12:20 it would have been me, jeff and jbroman as the 3 directors 2011-12-01T22:12:33 what happened to you and jbroman? 2011-12-01T22:12:33 but i think jbroman didn't want to do it because his affiliation with Google 2011-12-01T22:13:00 but i think you need at least 3 directors 2011-12-01T22:13:40 you, Jeff and Jeff's nearest relation ;) 2011-12-01T22:18:58 bmh: I went back to v4 w/attack at someone's request on irc 2011-12-01T22:19:07 let me know if I should revert later 2011-12-01T22:19:26 BenJackson: no prob. The issue is bad pairing, more than your bot 2011-12-01T22:19:37 yeah I don't know fi the tcp servers do any matchmaking at all 2011-12-01T22:19:41 and yours is not doing truskill 2011-12-01T22:19:54 grr... how do I enable that? 2011-12-01T22:20:02 good q for janzert and amstan 2011-12-01T22:20:07 probably just drop in the java package 2011-12-01T22:20:27 I've got a bot to write :) 2011-12-01T22:20:35 I really don't know anything about how the tcp server is setup to do ranking or matchmaking 2011-12-01T22:20:49 matchmaking is first come - first serve 2011-12-01T22:22:23 *** Antimony has joined #aichallenge 2011-12-01T22:50:03 *** JorgeB has joined #aichallenge 2011-12-01T23:02:56 *** bmh has quit IRC (Ping timeout: 265 seconds) 2011-12-01T23:03:01 *** dvladim has joined #aichallenge 2011-12-01T23:03:49 bots are getting way too good on tcp 2011-12-01T23:03:53 i've lost every single game 2011-12-01T23:04:01 wit h0 2011-12-01T23:06:44 *** Lithosphere has quit IRC (Ping timeout: 244 seconds) 2011-12-01T23:18:01 *** avdg has quit IRC (Quit: Leaving.) 2011-12-01T23:18:13 *** dvladim has quit IRC (Ping timeout: 240 seconds) 2011-12-01T23:20:08 *** avdg has joined #aichallenge 2011-12-01T23:21:24 *** u_ has quit IRC (Quit: u_) 2011-12-01T23:23:33 tcp server is definitely not for the faint hearted 2011-12-01T23:26:50 *** Jak_o_Shadows has joined #aichallenge 2011-12-01T23:43:56 *** ThiagoRRamos has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111104165243]) 2011-12-01T23:48:10 *** gd_ has joined #aichallenge 2011-12-01T23:55:36 *** gd_ has quit IRC (Quit: Page closed) 2011-12-01T23:56:35 *** delt0r_ has quit IRC (Ping timeout: 244 seconds)