2011-11-10T00:09:51 *** twymer has quit IRC (Ping timeout: 258 seconds) 2011-11-10T00:11:00 I think I figured out a way to reduce the storage necessary for the shortest path from any point to any other point on a 200x200 map from 3.2Gb to ~100Kb 2011-11-10T00:11:27 much smaller than 100Kb depending on the map 2011-11-10T00:12:45 and reduce the number of searchable nodes in a pathfinding algorithm from 40000 to ~200 2011-11-10T00:13:27 for long distance paths, and for shorter distance no pathfinding is even necessary 2011-11-10T00:16:02 *** Jak_o_Shadows1 has joined #aichallenge 2011-11-10T00:18:41 *** anon has joined #aichallenge 2011-11-10T00:19:11 *** Jak_o_Shadows has quit IRC (Ping timeout: 248 seconds) 2011-11-10T00:19:16 Does anyone know how to profile a Python3 bot (to debug timeout error)? 2011-11-10T00:24:46 *** Lintendo has quit IRC (Quit: Page closed) 2011-11-10T00:26:56 *** qqq has joined #aichallenge 2011-11-10T00:27:44 *** TheDigitalNinja has joined #aichallenge 2011-11-10T00:27:57 Does somebody have erlang starter package? 2011-11-10T00:30:19 the rules say multiple processes is prohibited, so it cannot use spawn in erlang, right? 2011-11-10T00:38:54 *** qqq has quit IRC (Quit: Page closed) 2011-11-10T00:38:55 *** Antimony has quit IRC (Ping timeout: 248 seconds) 2011-11-10T00:44:20 userjjb: are you going to tell us what your idea is or were you just gloating? ;) 2011-11-10T00:44:31 anon: can't you just use the normal python way? i.e. cProfile? 2011-11-10T00:50:14 *** Areks has joined #aichallenge 2011-11-10T00:53:30 bqf: I threw it our there since I know it's late so I figured if/when someone read what I wrote then they'd ask 2011-11-10T00:53:40 1 sec I have a good explanatory imagfe 2011-11-10T00:56:37 *** AndrewBC has quit IRC (Read error: Operation timed out) 2011-11-10T00:56:42 *** nickjohnson has quit IRC (Read error: Operation timed out) 2011-11-10T00:56:56 *** Sir_Ragnarok has quit IRC (Read error: Connection reset by peer) 2011-11-10T00:58:17 *** luizribeiro has quit IRC (Read error: Operation timed out) 2011-11-10T00:58:32 *** Sir_Ragnarok has joined #aichallenge 2011-11-10T00:59:54 *** AndrewBC has joined #aichallenge 2011-11-10T00:59:59 How to disable Cutoffs locally? 2011-11-10T01:01:25 what go version is the server running? 2011-11-10T01:02:15 *** antimatroid has quit IRC (Ping timeout: 252 seconds) 2011-11-10T01:02:28 bqf: here you go http://farm7.static.flickr.com/6032/6331267532_5572e2d071_b.jpg 2011-11-10T01:02:39 *** Sir_Ragnarok_ has joined #aichallenge 2011-11-10T01:03:01 *** antimatroid has joined #aichallenge 2011-11-10T01:03:17 rmmh: that can be found at http://aichallenge.org/starter_packages.php 2011-11-10T01:03:21 *** Sir_Ragnarok has quit IRC (Ping timeout: 252 seconds) 2011-11-10T01:03:22 *** Bluedgis has quit IRC (Ping timeout: 252 seconds) 2011-11-10T01:03:26 *** Bluedgis has joined #aichallenge 2011-11-10T01:03:34 hmm... is there any chance of getting it updated? 2011-11-10T01:04:03 there have been another 700 revisions since that version 2011-11-10T01:04:31 first response is 'nope', but I suppose there is some very slim chance but really it's pretty much just 'nope' 2011-11-10T01:04:59 welp 2011-11-10T01:05:20 *** luizribeiro has joined #aichallenge 2011-11-10T01:05:21 there's an automated tool that updates go programs to work with the latest release 2011-11-10T01:05:31 so it's not like contestants would have to spend time 2011-11-10T01:05:42 bqf: the only pathfinding you need to do/save is between regions, and generally the number of regions goes up as square root of map size, potentially much slower depending onhow "open" the map is 2011-11-10T01:06:00 yes, I know but I'm not willing to completely trust it to run 100% correct and automatically on the current submissions 2011-11-10T01:06:14 userjjb: That's neato 2011-11-10T01:06:26 do submissions get compiled more than once? 2011-11-10T01:06:35 and it's not so much contestants that I worry about with language upgrades it's the current submissions 2011-11-10T01:06:53 datachomper: thanks! I haven't any code to implement it yet, I thought of it at school today while I was reading about pathfinding 2011-11-10T01:07:03 yes, submissions get recompiled on each worker and when new workers come up 2011-11-10T01:07:13 ugh 2011-11-10T01:07:21 I guess I'll just backport my code when I submit 2011-11-10T01:08:41 userjjb: it's a small world. I've thought of that today too and have still no code about it. 2011-11-10T01:09:37 LouisMartin: yeah, I'm not naive enough to think it's the first time someone's thought of it, but I haven't read about it in any of the myriad pathfinding literature that I've read yet 2011-11-10T01:10:03 userjjb: How do you travel from a point in one region, to a point in a disconnected region? 2011-11-10T01:11:21 userjjb: yea, me neither. Some may have implemented this idea yet, it still will stay an "out-of-the-box" idea I'd think. A*Star, BFS, Minimax = 95% of all decent bot in the contest at the moment. 2011-11-10T01:11:40 *** JP_ has joined #aichallenge 2011-11-10T01:11:47 My understanding of graph theory is that a lot of the "pathfinding" algorithms are for a generic graph, whereas the type of graph that represents a grid with only NESW movement is relatively special, it's certainly "sparse" it's only 1/10000 connected as much as a fully dense graph, so I figured there should certainly be some specialized pathfinding algorithms possible 2011-11-10T01:13:04 Janzert: is there any way to tell the engine to not stop the game whatever happens? or as long as there's some enemy hills? 2011-11-10T01:13:25 datachomper: Well the idea I have in my head is treating each region as it's own "cell" and applying a conventional pathfinding algo on it 2011-11-10T01:13:47 *** Fandekasp has joined #aichallenge 2011-11-10T01:14:00 *** anon has quit IRC (Quit: Page closed) 2011-11-10T01:14:03 It reduces the number of nodes by so much that even storing all optimal paths from every sqaure to every other one should only take ~100Kb 2011-11-10T01:14:04 userjjb: that's how I see it too. 2011-11-10T01:14:17 LouisMartin: not that I've heard of but I've really only had limited experience with the engine code 2011-11-10T01:14:20 *** antimatroid has quit IRC (Ping timeout: 260 seconds) 2011-11-10T01:14:43 janzert: okay, thanks 2011-11-10T01:15:23 I think you could easily modify the game_over method in ants.py to do what you want but I may be missing something 2011-11-10T01:15:43 *** JP_ has quit IRC (Client Quit) 2011-11-10T01:16:27 LouisMartin: I imagine that if this method is doable I'll constrain ants to mostly stay in a given region with interregion travel only occuring to newly found regions, so the majority of ant motion won't even require pathfinding! 2011-11-10T01:16:36 *** JP_ has joined #aichallenge 2011-11-10T01:17:26 *** Jak_o_Shadows has joined #aichallenge 2011-11-10T01:17:26 janzert: indeed. I'll try 2011-11-10T01:17:42 LouisMartin: The trick is creating and filling in the regions as the knowledge of the map grows. 2011-11-10T01:17:48 I think 2011-11-10T01:18:33 *** dorisabayon has joined #aichallenge 2011-11-10T01:19:35 *** Jak_o_Shadows1 has quit IRC (Ping timeout: 260 seconds) 2011-11-10T01:20:29 *** antimatroid has joined #aichallenge 2011-11-10T01:21:07 userjjb: check your private msgs 2011-11-10T01:22:44 *** ltriant has quit IRC (Quit: Computer has gone to sleep) 2011-11-10T01:26:36 *** tobym has joined #aichallenge 2011-11-10T01:26:58 *** antimatroid has quit IRC (Ping timeout: 260 seconds) 2011-11-10T01:27:16 *** antimatroid has joined #aichallenge 2011-11-10T01:35:37 *** roflmao has quit IRC (Quit: Leaving.) 2011-11-10T01:48:29 *** Palmik has joined #aichallenge 2011-11-10T01:56:13 *** amstan has quit IRC (Ping timeout: 245 seconds) 2011-11-10T02:09:09 *** epicmonkey has joined #aichallenge 2011-11-10T02:10:24 Who should I contact if I found an error in the documentation on the website? 2011-11-10T02:11:08 *** datachomper has quit IRC (Quit: Leaving.) 2011-11-10T02:11:23 *** Mooloo has joined #aichallenge 2011-11-10T02:11:27 *** datachomper has joined #aichallenge 2011-11-10T02:14:30 *** JP_ has quit IRC (Ping timeout: 240 seconds) 2011-11-10T02:15:04 *** Flort has quit IRC (Ping timeout: 255 seconds) 2011-11-10T02:15:31 *** pairofdice has joined #aichallenge 2011-11-10T02:18:39 *** delt0r_ has quit IRC (Ping timeout: 258 seconds) 2011-11-10T02:22:09 *** AVAVT has joined #aichallenge 2011-11-10T02:23:15 *** epicmonkey has quit IRC (Ping timeout: 276 seconds) 2011-11-10T02:32:03 *** delt0r_ has joined #aichallenge 2011-11-10T02:32:57 *** JP_ has joined #aichallenge 2011-11-10T02:36:48 *** djr_ has joined #aichallenge 2011-11-10T02:41:37 *** mleise has joined #aichallenge 2011-11-10T02:44:42 *** freshAnt has quit IRC (Ping timeout: 265 seconds) 2011-11-10T02:47:11 *** contestbot_ has joined #aichallenge 2011-11-10T02:48:33 *** contestbot has quit IRC (Ping timeout: 258 seconds) 2011-11-10T02:50:22 *** ikaros has joined #aichallenge 2011-11-10T02:58:17 *** kilae has joined #aichallenge 2011-11-10T03:04:11 It is hazardous ant awareness week 2011-11-10T03:04:52 I just realized why my ants didn't care anymore 2011-11-10T03:05:37 I initialized the dictionary that keeps track of CombatAnts each loop >_< 2011-11-10T03:05:48 * pairofdice facepalms 2011-11-10T03:07:51 *** tobym has quit IRC (Remote host closed the connection) 2011-11-10T03:10:06 I also found the glory of .setdefault() 2011-11-10T03:19:54 *** JP_ has quit IRC (Quit: Leaving) 2011-11-10T03:20:33 *** waterbed has joined #aichallenge 2011-11-10T03:22:47 *** Blkt has joined #aichallenge 2011-11-10T03:23:25 *** aerique has joined #aichallenge 2011-11-10T03:24:23 *** mj41 has joined #aichallenge 2011-11-10T03:25:39 mmmm freedom 2011-11-10T03:27:42 I just figured out why all passages have to be at least 3 wide 2011-11-10T03:29:03 *** kaemo has quit IRC (Ping timeout: 260 seconds) 2011-11-10T03:29:05 on any given map 2011-11-10T03:31:34 :) 2011-11-10T03:32:24 we don't really need all passages to be 3 wide, but would prefer a 3x3 block to be able to traverse the map (except maybe the edges of walls) 2011-11-10T03:32:27 *** NotABug has joined #aichallenge 2011-11-10T03:33:26 Right, but if passages weren't at least 3 wide it would be possible to make an impenetrable wall of ants at the mouth of the passage 2011-11-10T03:33:50 You can attack over water so not really 2011-11-10T03:34:27 Also why would an antwall of 4 wide be any less impenetrable than 2 wide? 2011-11-10T03:35:01 userjjb: yeah but if you could divert ants around to the back of their block i wouldn't mind so much 2011-11-10T03:35:03 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-11-10T03:35:20 hence having a 3x3 block being able to traverse the map but not necessarily go through each corridor 2011-11-10T03:35:49 pairofdice: water doesn't matter either way imagine a "hat" line of 4 ants sitting atop a 2-wide corridor 2011-11-10T03:35:53 pairofdice: battle resolution allows blocking of 1 and 2 width corridors 2011-11-10T03:36:14 pairofdice: you can only get 2 ants within range of 4 ants, they would always lose 2011-11-10T03:36:15 It allows for mutual destruction 2011-11-10T03:36:39 You can move an antblock of 4 as asingle unit 2011-11-10T03:36:56 but the back two would be out of range 2011-11-10T03:37:43 antimatroid: I don't think it does 2011-11-10T03:38:10 can't say i've checked with the current attack radius but i assure you it was the case previously 2011-11-10T03:38:18 we did extensive testing on possible battle options 2011-11-10T03:38:35 attackradius2=5 right? 2011-11-10T03:38:47 not sure, been busy with exams etc. 2011-11-10T03:38:59 just got free earlier 2011-11-10T03:39:21 I made a handy excel sheet that visualizes number of ants within range of a square 2011-11-10T03:39:29 Would be possible if friendly ants could hold hands and channel their chi to the one ant blocking the way 2011-11-10T03:39:34 so I've been looking at various formations 2011-11-10T03:40:02 But now you can justdestry a block by expending an equal amount of ants at it most of the time 2011-11-10T03:40:10 *** kiv has joined #aichallenge 2011-11-10T03:40:14 right 2011-11-10T03:40:16 pairofdice: it's easy to test, make a bot that just always moves north and a bot that does nothing. Then create a map with a column of ants n wide in the middle and a block of ants at least n+2 wide at the top 2011-11-10T03:40:35 for n < 3 you should find the column just gets wiped out with no compensation 2011-11-10T03:40:49 That's weird 2011-11-10T03:41:02 oh you have to run the map with the -scenario option of course 2011-11-10T03:41:28 I don't think the game would be broken per se with 2 wide corridors, but ant walls would be a *very* strong strategy 2011-11-10T03:41:55 pretty much mandatory against anyone else who uses them 2011-11-10T03:42:31 *** replore has quit IRC (Remote host closed the connection) 2011-11-10T03:42:56 Well okay sure, but to my knowledge, uniformly wide corridors don't allow one block to be n+2 while the other can be only n 2011-11-10T03:43:31 except at entrances to "rooms" or at "T" junctions 2011-11-10T03:43:36 the wall is placed at the end outside of the corridor 2011-11-10T03:44:18 Yeah... 2011-11-10T03:44:22 also, with the current "hairyness" in maze maps, if a corridor was 2 wide it might be possible to find a spot in a corridor where a 1 wide divot occurs on either side of the corridor 2011-11-10T03:44:23 okay 2011-11-10T03:46:54 Because the ant attack radius makes a block 3 at sides with a dent in the corners 2011-11-10T03:47:20 Well the dent doesn't matter just the size 2011-11-10T03:48:52 right, if you're familiar with elctromagnetic physics, it's akin to the electric field at a certain distance away from an "infinite" line of charge 2011-11-10T03:48:59 *** NotABug has quit IRC (Ping timeout: 265 seconds) 2011-11-10T03:49:40 *** mviel has joined #aichallenge 2011-11-10T03:49:51 A bit odd analogy but I'll take it :) 2011-11-10T03:49:57 *** retybok has joined #aichallenge 2011-11-10T03:53:12 I haven't watched a whole lot of battles, but I haven't really seen anybodies ants moving around in a particular formation for battle purposes, do people do that? 2011-11-10T03:53:17 *** theskumar has quit IRC (Quit: bye!) 2011-11-10T03:53:48 Not really, no 2011-11-10T03:54:02 I thought of using a "powerblock" but it just seems inefficient 2011-11-10T03:55:26 Right, well the thing is that a formation is always weak at it's corners or edges, so it's easy to defeat if you are trying to, but if the opponents ants don't bother moving in at least 2's it's easy to kill individual ants 2011-11-10T03:55:41 *** Septimus has joined #aichallenge 2011-11-10T03:55:54 or at the very least make them avoid/run away 2011-11-10T03:56:07 Would be funny though 2011-11-10T03:56:39 I'd imagine people are fleching out the basics, and as bots get more mature greater attention will be paid towards it 2011-11-10T03:56:46 fleshing* 2011-11-10T03:57:10 *** LouisMartin has quit IRC (Quit: Page closed) 2011-11-10T03:58:14 *** Ionic_Groove has quit IRC (Ping timeout: 258 seconds) 2011-11-10T03:58:17 Things is, if people aren't doing that, it seems better to avoid doing it yourself, lest an arms race develops. Better to hold back the weapon until the actual competition and then drop the "nuke" that no one has a defense to 2011-11-10T03:59:23 *** sigh has joined #aichallenge 2011-11-10T04:01:43 Yeah, the general level of the bots will probably sharply rise towards the end 2011-11-10T04:02:31 That should happen regardless of 'hoarding' though, too 2011-11-10T04:06:51 *** mviel_ has joined #aichallenge 2011-11-10T04:09:21 *** djr_ has joined #aichallenge 2011-11-10T04:09:28 *** djr_ has quit IRC (Read error: Connection reset by peer) 2011-11-10T04:10:43 *** mviel has quit IRC (Ping timeout: 256 seconds) 2011-11-10T04:15:27 *** Fandekasp has quit IRC (Ping timeout: 248 seconds) 2011-11-10T04:15:42 *** dorisabayon has quit IRC (Ping timeout: 276 seconds) 2011-11-10T04:17:10 *** ikaros has joined #aichallenge 2011-11-10T04:19:02 userjjb: watching the final matches will be fun for sure :) 2011-11-10T04:19:31 anyone know what the "extended different" suffix in ending reasons are? 2011-11-10T04:39:32 *** NotABug has joined #aichallenge 2011-11-10T04:46:48 *** rajanaresh has left #aichallenge 2011-11-10T04:49:31 *** eccentric has quit IRC (Ping timeout: 260 seconds) 2011-11-10T04:52:06 *** antimatroid has quit IRC (Ping timeout: 240 seconds) 2011-11-10T04:53:09 *** waterbed has quit IRC (Quit: Page closed) 2011-11-10T04:56:12 AVAVT: presumambly some sort of exception case, the finally else in a if else if chain that is put there for safety 2011-11-10T04:57:30 oic 2011-11-10T04:58:12 if you look at game stats it's very rare 2011-11-10T04:58:18 only something like 10 games 2011-11-10T04:58:40 *** antimatroid has joined #aichallenge 2011-11-10T05:04:45 *** DigitalNinja has joined #aichallenge 2011-11-10T05:05:42 *** TheDigitalNinja has quit IRC (Ping timeout: 258 seconds) 2011-11-10T05:07:39 *** antimatroid1 has joined #aichallenge 2011-11-10T05:07:42 *** antimatroid has quit IRC (Ping timeout: 240 seconds) 2011-11-10T05:12:03 *** antimatroid1 has quit IRC (Ping timeout: 245 seconds) 2011-11-10T05:20:08 *** Cyndre has quit IRC (Quit: Leaving) 2011-11-10T05:26:21 *** Zaphus has joined #aichallenge 2011-11-10T05:29:22 *** liberforce has joined #aichallenge 2011-11-10T05:32:52 Holy gaussian distribution, progres happened! 2011-11-10T05:33:33 Now my ants fall back untill they find a friendly. 2011-11-10T05:34:58 Sort of 2011-11-10T05:39:14 Why is map wrapping still haunting me 2011-11-10T05:44:26 Oh, that was just the older bot *facepalm* 2011-11-10T05:44:32 *** AVAVT has quit IRC (Quit: Page closed) 2011-11-10T05:52:21 *** mviel__ has joined #aichallenge 2011-11-10T05:56:06 *** mviel_ has quit IRC (Ping timeout: 256 seconds) 2011-11-10T05:56:28 *** DigitalNinja has quit IRC (Remote host closed the connection) 2011-11-10T05:58:57 *** mcstar has joined #aichallenge 2011-11-10T05:59:59 *** antimatroid has joined #aichallenge 2011-11-10T06:07:53 *** praveen has joined #aichallenge 2011-11-10T06:08:04 (a late) good morning everyone 2011-11-10T06:09:17 9992 2011-11-10T06:09:57 *** NoxiaZ^ has joined #aichallenge 2011-11-10T06:10:01 *** chris__0076 has joined #aichallenge 2011-11-10T06:11:43 *** Chris_0076 has quit IRC (Ping timeout: 248 seconds) 2011-11-10T06:12:32 My previous bot peaked at about 430, this one should do a bit better. But who knows how much inflation has happened 2011-11-10T06:12:57 what's the bot name ? 2011-11-10T06:13:33 *** ericlavigne has joined #aichallenge 2011-11-10T06:13:53 MyBot.py3 ;) 2011-11-10T06:14:03 lol 2011-11-10T06:14:13 which server is that on ? 2011-11-10T06:14:27 Just on the official 2011-11-10T06:17:03 *** thakidd has joined #aichallenge 2011-11-10T06:17:22 *** Jak_o_Shadows1 has joined #aichallenge 2011-11-10T06:17:22 it takes a fair bit of time to see your ranking on the official server... I uploaded one two days ago and it's managed 7 games so far 2011-11-10T06:17:59 I would like my bots to print messages to my while running a game on my own computer, but it seems like playgame.py captures both standard-input and standard-error streams, and even blocks writing to files. What is a good way to do logging? 2011-11-10T06:18:29 hello im kind of a newb programmer and i was hoping someone can give me some help i would like to prevent the ants from oscillating any ideas? 2011-11-10T06:19:04 ericlavigne, use -e option for playgame.py 2011-11-10T06:19:19 what does -e do ? 2011-11-10T06:19:26 redirect stderr 2011-11-10T06:19:41 i wanted to implore a last turn order list type idea but not sure how to institute it 2011-11-10T06:19:46 but writing to files should work anyways 2011-11-10T06:19:48 ikaros: I was using -e from the beginning. Adding -E made some error logs appear in the gamelog directory, but those files are always empty :-( 2011-11-10T06:20:09 I was very surprised when writing to file didn't work. 2011-11-10T06:20:31 *** Jak_o_Shadows has quit IRC (Ping timeout: 260 seconds) 2011-11-10T06:20:38 I know how to redirect stderr on Linux. I am on Windows right now. Still doable? 2011-11-10T06:20:46 thakidd: store places you have been to and try to avoid them 2011-11-10T06:20:52 no idea about windows :) 2011-11-10T06:21:04 thakidd: using a global list means you don't have to track individual ants 2011-11-10T06:22:34 pairofdice: usually when resubmitting an improved bot I end up where I was before :) pretty frustrating! 2011-11-10T06:22:44 retybok: right but how do i add to the list and clear it effectively 2011-11-10T06:23:05 ericlavigne, have you perhaps looked for your log files in the wring directory? 2011-11-10T06:23:19 thakidd: I'm not going to tell you everything ;-) but that simple idea is enough to get ranked 1500th 2011-11-10T06:23:21 because writing to file isnt blocked for sure 2011-11-10T06:24:12 retybok: the problem i have is i know some very effective techniques but im unclear on the code to implement them 2011-11-10T06:24:44 ikaros, Maybe there are logs in more than one directory? The game creates a folder called game_logs in same directory as I run it from. The error files in that directory are empty. When I write to a file myself, I use an absolute path, so I'm certain that I look in the right place for that. 2011-11-10T06:24:57 hmm 2011-11-10T06:24:58 retybok: i just need a helping hand with the java code itself not the ideas 2011-11-10T06:25:32 thakidd: Have you used the tutorial? It includes example Java code for each step. 2011-11-10T06:25:33 the game_logs folder is something different .. i think its for replays or so 2011-11-10T06:26:02 ikaros: Then maybe I am looking in the wrong place. Where would you expect error logs for my bots to go? 2011-11-10T06:26:04 but you should have some kind of io errors if writing to disk fails 2011-11-10T06:26:20 well you say you used an absolute path so.. there :) 2011-11-10T06:27:04 retybok: i have but i want to get away from the manhattan algorithm and use a* along with my no backsies idea just not sure how to populate the list each turn and then clear it out 2011-11-10T06:27:09 ikaros, checked the absolute path. That io function works fine when invoked from outside the game. 2011-11-10T06:27:27 check if writing to disk fails in your source and just terminate the bot then 2011-11-10T06:28:00 so at least you see if the function fails or if its something different 2011-11-10T06:28:57 ikaros, good point. I can read the file back immediately after a write. If contents aren't what I expect, throw an exception. 2011-11-10T06:29:09 *** eccentric has joined #aichallenge 2011-11-10T06:29:49 ericlavigne:i have but i want to get away from the manhattan algorithm and use a* along with my no backsies idea just not sure how to populate the list each turn and then clear it out 2011-11-10T06:29:58 *** caution has joined #aichallenge 2011-11-10T06:31:01 thakidd, check out the HashSet class. It has methods for adding items, checking if an item has already been added, and clearing the list. 2011-11-10T06:33:13 No players on the tcp server? 2011-11-10T06:33:34 I'm there (fluxid server that is) - but no games starting 2011-11-10T06:34:28 it does that sometimes 2011-11-10T06:34:45 there's another TCP server you could try 2011-11-10T06:36:31 I can't find it, link me if you know it 2011-11-10T06:38:16 *** Jak_o_Shadows1 has quit IRC (Remote host closed the connection) 2011-11-10T06:44:01 there is one here: 213.88.39.97 thats running right now 2011-11-10T06:46:44 *** thakidd has quit IRC (Quit: IRC webchat at http://irc2go.com/) 2011-11-10T06:47:00 358 players ahead of me on the main contest server... hoping to get one last game in before I go to bed, but it's looking slim 2011-11-10T06:50:42 *** olexs has joined #aichallenge 2011-11-10T06:53:24 *** grwip has joined #aichallenge 2011-11-10T06:54:54 *** NoxiaZ^ has quit IRC (Ping timeout: 240 seconds) 2011-11-10T06:55:00 where can I learn how to use the TCP server? 2011-11-10T06:55:51 ericlavigne: I think there are some instructions on the site itself 2011-11-10T06:55:55 it's pretty easy 2011-11-10T06:55:56 http://ants.fluxid.pl/howto 2011-11-10T06:56:58 has anyone implemented Jump Point Search yet? 2011-11-10T06:57:18 *** ikaros has quit IRC (Ping timeout: 240 seconds) 2011-11-10T06:57:23 looks like there hasnt been a game on fluxid for 30 minutes or so 2011-11-10T06:57:34 Zaphus: Minthos : it is down apparently 2011-11-10T06:57:40 my bot gets invalid input from the server 2011-11-10T06:57:56 Fluxid: it seems that your server is down 2011-11-10T06:59:02 Looks like each player gets a game every few minutes on that server, this will make a huge difference in testing :-) 2011-11-10T06:59:05 caution: I don't have speed problems, so this type of algorithm is overkill for my bot 2011-11-10T06:59:34 ericlavigne: you get a lot more game because your bot runs on your machine 2011-11-10T06:59:42 heh, lucky you, only a few of my ants can afford to pathfind 2011-11-10T06:59:42 so the server does less 2011-11-10T07:00:22 Fluxid: may I recommend setting up an auto server restart on inactivity in lieu of a better fix 2011-11-10T07:01:04 i wonder how much slower is python (interpreted) compared to some of the compiled languages? 2011-11-10T07:01:25 in the top 100, it's something like 8 python vs 35 C++ 2011-11-10T07:02:24 there are speed comparisons between languages that can be found online 2011-11-10T07:02:28 you can use python, but it means you have to be very careful 2011-11-10T07:02:32 the sample python bots seem to take ages to process their turns, at least compared to my C bot 2011-11-10T07:02:34 thats probably the equivalent ratio of peoples experience with the languages as well 2011-11-10T07:02:36 JanneP_ if you're writing loops where a small amount of work is done in each loop, python is at least 100 times slower. Difference can be a lot smaller if you can do most work in well maintained libraries. 2011-11-10T07:02:51 I use Go, I've found it's a good tradeoff 2011-11-10T07:03:10 do you work for Google? 2011-11-10T07:03:16 nope 2011-11-10T07:03:27 I just took the challenge as an opportunity to learn the language 2011-11-10T07:03:48 what's your preferred language and how are you finding Go in comparison? 2011-11-10T07:06:17 I usually use python. Go is not as expressive, but it's pretty good and it's a lot faster 2011-11-10T07:06:32 The tools shipped with the language are pretty good (the profiler for example) 2011-11-10T07:07:03 And the static typing with type inference is actually a pretty good thing, ie. it catches a few errors that take time to debug in python 2011-11-10T07:07:29 retybok: have you had a look at scala? 2011-11-10T07:07:30 I don't like the look of the syntax 2011-11-10T07:07:37 *** g0llum has joined #aichallenge 2011-11-10T07:07:41 restarted server, Zaphus retybok 2011-11-10T07:07:43 caution: that's really a poor argument for rejecting a language 2011-11-10T07:07:48 caution: it does not crash cleanly 2011-11-10T07:07:52 syntax is something you get used to very quickly 2011-11-10T07:08:00 Fluxid: thanks! 2011-11-10T07:08:09 Fluxid: separate monitoring process 2011-11-10T07:08:13 jix: no, I don't know scala 2011-11-10T07:08:31 *** xbelt has joined #aichallenge 2011-11-10T07:08:50 Fluxid, yep, I'm in a game now 2011-11-10T07:08:58 caution: i have more important tasks atm :) 2011-11-10T07:09:12 retybok: it also uses static typing with type inference 2011-11-10T07:09:19 lol, I have more important sleep to deal with! 2011-11-10T07:09:50 jix: it looks like a decent language. 2011-11-10T07:10:09 retybok: yeah that's the language I decided to learn for this contest and I'm pretty happy with it (apart from the rather longish compile times) 2011-11-10T07:10:55 jix: Go has a very lightweight feeling: extremely fast compilation, instant program start (since no JVM to start), etc. 2011-11-10T07:11:30 @rank test 2011-11-10T07:11:30 caution: Error: The command "rank" is available in the ChannelStats and Factoids plugins. Please specify the plugin whose command you wish to call by using its name as a command before "rank". 2011-11-10T07:21:40 *** Zaphus has quit IRC (Quit: Page closed) 2011-11-10T07:24:17 Anyone ever use John the Ripper to crack a SHA1 hashed password? 2011-11-10T07:24:18 *** t2027 has joined #aichallenge 2011-11-10T07:25:33 I think I did a while ago 2011-11-10T07:28:07 How long did it take you? 2011-11-10T07:28:41 I've been hashing for 78 hours now :( 2011-11-10T07:29:03 I think I let it run the whole night and had results the next morning 2011-11-10T07:29:44 but this was really easy passwords 2011-11-10T07:31:24 *** mj41 has quit IRC (Ping timeout: 259 seconds) 2011-11-10T07:31:32 I'm just hoping it's not longer than 8 characters 2011-11-10T07:38:13 *** mj41 has joined #aichallenge 2011-11-10T07:44:31 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Ping timeout: 248 seconds) 2011-11-10T07:54:08 offline attacks like that only work on easy passwords 2011-11-10T07:54:16 and not typically on sha1 2011-11-10T07:54:24 the rainbow tables are too big 2011-11-10T07:54:34 @rainbow 10000 total users @ 12:53 2011-11-10T07:54:34 mcstar: 10000 total users @ 12:53 2011-11-10T07:54:44 lol 2011-11-10T07:54:45 haha 2011-11-10T07:54:51 Where are the unicorns 2011-11-10T07:54:57 !roulette 2011-11-10T07:55:01 @roulette 2011-11-10T07:55:01 delt0r: *click* 2011-11-10T07:55:11 @rainbow congrats organizers, goood job 2011-11-10T07:55:11 mcstar: congrats organizers, goood job 2011-11-10T07:55:35 @rainbow of course every single one of them is asleep 2011-11-10T07:55:35 mcstar: of course every single one of them is asleep 2011-11-10T07:56:10 but its only just after lunch ;) 2011-11-10T07:56:31 *** Insti has joined #aichallenge 2011-11-10T07:57:08 @rainbow rainbow rainbow rainbow! 2011-11-10T07:57:08 antimatroid: rainbow rainbow rainbow! 2011-11-10T07:57:23 mcstar: i'm free :D 2011-11-10T07:57:54 were you imprisoned? 2011-11-10T07:58:33 probably in a maze :-) 2011-11-10T07:58:44 *** NoxiaZ^ has joined #aichallenge 2011-11-10T07:58:48 no i did that to you :P 2011-11-10T07:58:55 mcstar: uni is all done 2011-11-10T07:59:21 technically i'm now unemployed, although i'll keep uniing next year if i get funding 2011-11-10T08:00:06 antimatroid: glad to hear it, if i were there i would throw a party 2011-11-10T08:00:23 i'm riding across the state starting sunday 2011-11-10T08:00:26 that's my present to myself aha 2011-11-10T08:00:30 whoa 2011-11-10T08:00:47 why are you punishing yourself? 2011-11-10T08:00:56 with a bicycle ride? 2011-11-10T08:00:59 yeah 2011-11-10T08:01:13 friend and i are getting a lift up to the north end and riding back along trails 2011-11-10T08:01:19 shall be sweet 2011-11-10T08:01:19 antimatroid: you must have some klingon blood in you 2011-11-10T08:01:30 i'm australia, we harden the f up :) 2011-11-10T08:01:35 australian* 2011-11-10T08:01:50 good an ya mate 2011-11-10T08:01:57 *** ericlavigne has quit IRC (Quit: Page closed) 2011-11-10T08:02:05 :p 2011-11-10T08:03:43 *** bergmark has quit IRC (Ping timeout: 248 seconds) 2011-11-10T08:09:38 Stefan 2011-11-10T08:11:46 Is it just me or don't the demos on golang.oeg work :p 2011-11-10T08:11:50 org 2011-11-10T08:12:44 Though looking at the installation instructions it's removed as an option 2011-11-10T08:17:43 *** Savaron has joined #aichallenge 2011-11-10T08:18:00 *** mrNub has joined #aichallenge 2011-11-10T08:19:32 delt0r: I'm not using rainbow tables as far as I know, I hope not because it's a salted hash 2011-11-10T08:20:27 *** HaraKiri has joined #aichallenge 2011-11-10T08:21:47 userjjb you need to rent an hour on an aws gpu instance http://stacksmashing.net/2010/11/15/cracking-in-the-cloud-amazons-new-ec2-gpu-instances/ 2011-11-10T08:22:17 asssuming I could enumerate all possible battle scenarios for 10x10 and generate a list with best battle behavior that I save to the file 2011-11-10T08:22:31 it would be allowed to upload that file with my bot and read from that file? 2011-11-10T08:22:54 *** Antimony has joined #aichallenge 2011-11-10T08:23:00 Sure 2011-11-10T08:23:36 But that's assuming a lot :) 2011-11-10T08:23:39 what do you think of the idea? 2011-11-10T08:23:41 I think it might be a pretty big file... 2011-11-10T08:23:51 Yeah 2011-11-10T08:24:51 Savaron: as long as it's under 2MB :) 2011-11-10T08:25:03 "If you need to submit a bigger file for some reason, that's fine. Post on the forums and we'll work something out." 2011-11-10T08:25:37 Savaron: I asked basically the same question yesterday 2011-11-10T08:25:44 :) 2011-11-10T08:25:44 So you'll have something to upload in a thousand years or so 2011-11-10T08:25:46 i saw that 2011-11-10T08:26:03 you are still capped by the startup time in terms of reading the file. ofc you could presumably mmap it if you were smart about constructing it 2011-11-10T08:28:48 *** ikaros has joined #aichallenge 2011-11-10T08:29:19 Savaron: Some programming languages offer compile time function evaluation. So you don't need to upload the list, but let the compiler generate it ;) 2011-11-10T08:29:53 *** dwins has joined #aichallenge 2011-11-10T08:30:00 unfortunately i'm a c# guy ;) 2011-11-10T08:30:54 and most of the time I have no clue what I'm doing, so thats just a mind game ;) 2011-11-10T08:31:42 Savaron: What do you mean battle scenarios? Ant positions prior to combat? And how far "prior" 2011-11-10T08:31:51 hmm, (10 players + 1 unoccupied) ^ 10x10 squares ... seems to be a bad idea 2011-11-10T08:32:17 mleise: I'm not sure the contest organizers would appreciate your compilation taking days :) 2011-11-10T08:32:22 my calculator says these are 137806123398222701841183371720896367762643312000384664331464775521549852095523076769401159497458526446001 possibilities 2011-11-10T08:32:32 relevant would be only 4 cases (ant, enemy, land, water) 2011-11-10T08:32:59 Savaron: if there are two different ennemies, it changes things a lot 2011-11-10T08:33:02 right, because 1606938044258990275541962092341162602522202993782792835301376 are so much better 2011-11-10T08:33:48 mleise: you can divide that a bit it you take symmetries into account 2011-11-10T08:33:54 it's still a big number :) 2011-11-10T08:34:29 I'd say 3x3 squares is the limit of what is feasible to precalculate 2011-11-10T08:34:33 lol 2011-11-10T08:34:35 10x10 2011-11-10T08:34:37 lol 2011-11-10T08:37:03 at least we have talked about it ;) 2011-11-10T08:39:14 Savaron: considered it couple of months back, i laughed at myself when i made a similar back-of-the-evelope calculation 2011-11-10T08:41:36 *** Antimony has quit IRC (Remote host closed the connection) 2011-11-10T08:41:46 the state space is huge :D 2011-11-10T08:42:01 compared to tron at least :p 2011-11-10T08:43:30 *** yoden has joined #aichallenge 2011-11-10T08:43:44 *** DeGi has joined #aichallenge 2011-11-10T08:44:37 antimatroid: at least no one is tempted by a global minimax :) 2011-11-10T08:45:05 it's still there with the battles 2011-11-10T08:45:13 minimax that is 2011-11-10T08:45:28 sure, but it's not really the same thing if it's not global 2011-11-10T08:45:43 plus you can't go very deep, even locally 2011-11-10T08:47:02 no, i'd be surprised if people are doing more than one layer 2011-11-10T08:48:25 *** willvarfar has joined #aichallenge 2011-11-10T08:48:25 antimatroid: be surprised 2011-11-10T08:48:32 with good pruning 2 is possible 2011-11-10T08:48:43 with hand-optimized assembly code 2011-11-10T08:48:44 for how many ants? 2011-11-10T08:48:54 well, couple 2011-11-10T08:48:57 3-4 2011-11-10T08:49:06 mcstar: does it really make your bot stronger? 2011-11-10T08:49:08 I've got the normal starter pack java visualiser; does it support the fancy 'v' lineart stuff? (can't get it to) 2011-11-10T08:49:24 retybok: dont know, im refining 1-ply now 2011-11-10T08:49:29 well, not now but very soon 2011-11-10T08:49:41 im thinking on a project euler problem now 2011-11-10T08:49:46 damn, maybe i should start writing my bot before leaving sunday 2011-11-10T08:50:08 *** bergmark has joined #aichallenge 2011-11-10T08:52:44 *** peyton has joined #aichallenge 2011-11-10T08:53:10 mcstar: what's your username on the site? 2011-11-10T08:53:47 antimatroid 2011-11-10T08:53:54 no, thats you 2011-11-10T08:54:06 agentsmith 2011-11-10T08:54:16 but dont take that bot too seriously 2011-11-10T08:55:06 *** twymer has joined #aichallenge 2011-11-10T08:55:10 how are you pruning? 2011-11-10T08:55:13 Well it's not like precalculating some things would be a bad idea. But im not sure what are good candidates 2011-11-10T08:55:18 just with minimax? and at what level? 2011-11-10T08:55:34 like are you pruning before you've even considered all moves for your ants? 2011-11-10T08:56:07 antimatroid: im busy now 2011-11-10T08:56:14 but later we may talk 2011-11-10T08:56:29 I precalculated some normal distributions for my bot, that's it 2011-11-10T08:56:33 anyway, im debugging and polishing it 2011-11-10T08:56:37 mcstar: what's the euler problem :P 2011-11-10T08:56:42 later 2011-11-10T08:57:02 antimatroid: google project euler 2011-11-10T08:57:11 awww i wanted to race :( 2011-11-10T08:57:17 pairofdice: i meant the problem number 2011-11-10T08:57:24 *** yoden has quit IRC (Quit: Leaving.) 2011-11-10T08:57:28 Yeah, I got that after I pressed enter 2011-11-10T08:58:31 I've averaged one or two eulers a month >_< 2011-11-10T08:58:54 *** yoden has joined #aichallenge 2011-11-10T08:59:48 i haven't done any in ages 2011-11-10T08:59:55 i've done 56 though 2011-11-10T09:00:13 So if you're careful about eliminating redundant translations, rotations, reflections, and enemy identity invariances; and are stingy with how you store the data you could precalculate for 2 enemies on a 5x5 and it would take only ~234.6 terabytes of space 2011-11-10T09:00:29 :D 2011-11-10T09:01:19 if you take 2x2 simultaneous games where each player can totally order the pure strategy space there are 144 ordinal equivalence classes 2011-11-10T09:01:25 games are ridiculous 2011-11-10T09:01:50 Nah, just some approaches to games 2011-11-10T09:01:53 You could reduce the space by a bit if you factor in that not all states are achievable game configurations 2011-11-10T09:01:57 *** xbelt has quit IRC (Ping timeout: 258 seconds) 2011-11-10T09:02:19 but that's worthy of a *very* rainy day 2011-11-10T09:02:29 does anyone here know what a normal form game is? 2011-11-10T09:02:46 wikipedia does 2011-11-10T09:02:52 i know what they are 2011-11-10T09:03:06 i did my honoures thesis on them and made some pretty cool symmetric game examples 2011-11-10T09:03:40 i'm lazy and linking you to my entire talk, but go right to the examples at the end.. 2011-11-10T09:03:59 https://docs.google.com/open?id=0B5joEFt1EpwZMjBiNmFhNzktMzExMi00ZDI1LWI3MjUtMjNhNDJmNTViODRm 2011-11-10T09:04:56 the very last one is a fairly trivial game, but it's the only example in that class that i've been able to find 2011-11-10T09:05:38 *** sigh has quit IRC (Remote host closed the connection) 2011-11-10T09:08:05 Ahh, finally my bot can consistently beat my old bot 2011-11-10T09:09:13 Still some peculiarities 2011-11-10T09:11:31 But now I know what to do about those 2011-11-10T09:22:12 HAHA OMG, I'm playing around with something in a test program. I wanted to see if it was reading properly so I had it output to the console. When I run it it fills it with weird patterns of smiley faces! 2011-11-10T09:24:29 *** xbelt has joined #aichallenge 2011-11-10T09:25:12 :) 2011-11-10T09:26:25 no love for my symmetric games? :( 2011-11-10T09:26:54 *** twymer has quit IRC (Ping timeout: 240 seconds) 2011-11-10T09:27:47 pairofdice: oh, also the definition of a symmetric game given on wikipedia is wrong :P 2011-11-10T09:28:12 Does mirror-go qualify? 2011-11-10T09:28:28 ? 2011-11-10T09:28:37 *** Areks has quit IRC (Ping timeout: 240 seconds) 2011-11-10T09:28:46 As a symmetric game ;) 2011-11-10T09:29:03 *** xbelt has quit IRC (Ping timeout: 248 seconds) 2011-11-10T09:29:08 *** xbelt has joined #aichallenge 2011-11-10T09:29:16 Ah well, I think my code would get noticeably faster with the elimination of some for-loops 2011-11-10T09:29:47 i mean symmetric normal form games 2011-11-10T09:30:29 Reading that paper's too hard for this high-school drop-out :> 2011-11-10T09:31:06 you just need to look at the last 4 slides :P 2011-11-10T09:32:23 antimatroid - and over my head 20 years out from my last abstract algebra class 2011-11-10T09:33:04 bugnuts2: there isn't that much algebra once you make it past the maths stuff 2011-11-10T09:33:19 and i really just want like the groupoids to note when composition is closed and inverses exist 2011-11-10T09:33:30 err, when composition exists 2011-11-10T09:33:33 not necessarily closed 2011-11-10T09:33:41 *** tobym has joined #aichallenge 2011-11-10T09:34:58 and a matching is a simple concept, suppose you have A1 = {a,b}, A2 = {c,d}, A3 = {e, f}, a matching might be M = {(a,d,e), (b,c,f)}, ie. it's a way of matching up the elements in the sets 2011-11-10T09:35:10 i have no idea what people call those, i just defined them myself :\ 2011-11-10T09:38:24 antimatroid: I'll take a look later 2011-11-10T09:38:36 I like exotic math 2011-11-10T09:38:49 retybok: it takes like 10 seconds to look at the games at the end 2011-11-10T09:38:58 I'm at work :-/ 2011-11-10T09:39:01 *** _ud <_ud!c3dabef0@gateway/web/freenode/ip.195.218.190.240> has joined #aichallenge 2011-11-10T09:39:12 fair call 2011-11-10T09:39:26 funny time :P it's 1 40am 2011-11-10T09:39:35 antimatroid: your last name is ham? awesome! 2011-11-10T09:39:42 *** amstan has joined #aichallenge 2011-11-10T09:39:42 *** ChanServ sets mode: +o amstan 2011-11-10T09:40:07 Unless he's vegetarian 2011-11-10T09:41:19 not vegetarian and yes aha 2011-11-10T09:42:33 antimatroid: I think we should get rid of timezones alltogether :) 2011-11-10T09:42:41 resubmitted, and found activate/deactivate buttons in my profile. what's that ? 2011-11-10T09:42:56 I don't know if people would get used to the fact that the sun shines at midnight in half of the globe 2011-11-10T09:43:22 antimatroid: how did you make your mapgen to keep bases connected? 2011-11-10T09:44:00 amstan: not sure :\ 2011-11-10T09:44:08 i'll have a look and see if i can remember 2011-11-10T09:45:12 here's a sample: http://paste.aichallenge.org/4zLfo/ 2011-11-10T09:45:22 *** mviel_ has joined #aichallenge 2011-11-10T09:49:07 *** mviel__ has quit IRC (Ping timeout: 255 seconds) 2011-11-10T09:51:45 *** Ionic_Groove has joined #aichallenge 2011-11-10T09:52:14 amstan: what's wrong with the sample? 2011-11-10T09:52:31 antimatroid: nothing, showing you what i've done with my mapgen 2011-11-10T09:52:39 it looks good :) 2011-11-10T09:53:02 what happens when you make a fully sized map? :P 2011-11-10T09:54:30 antimatroid: define full sized 2011-11-10T09:54:40 like 100x100 would do 2011-11-10T09:54:42 ok 2011-11-10T09:54:43 i want to see that 2011-11-10T09:55:01 1000x1000 <- for real ants :D 2011-11-10T09:55:35 http://paste.aichallenge.org/UPF7f/ 2011-11-10T09:56:13 *** ifdef has joined #aichallenge 2011-11-10T09:56:55 I predict long queues in the hive for that :) 2011-11-10T09:57:15 amstan: that's awesome 2011-11-10T09:57:16 yeah, i still need to get a clearing for the immediate area of the hive 2011-11-10T09:57:28 but if i get the connectivity test, i think it'll solve that too 2011-11-10T09:57:43 *** twymer has joined #aichallenge 2011-11-10T09:57:52 another one: http://paste.aichallenge.org/JvhNP/ 2011-11-10T09:57:59 one way to test connectivity is a bfs from one ant hill to any other 2011-11-10T09:58:06 if two are connected they're all connected 2011-11-10T09:58:06 *** twymer has left #aichallenge 2011-11-10T09:58:10 *** twymer has joined #aichallenge 2011-11-10T09:58:16 that is if there's a unique hill for each player 2011-11-10T09:58:22 then add extra hills after 2011-11-10T09:58:34 antimatroid: it's not only about connectivity, i just want to make sure there's tunnels between hills 2011-11-10T09:58:50 define tunnel 2011-11-10T09:58:56 a path between hills 2011-11-10T09:59:05 ? 2011-11-10T09:59:10 i want to do it without needing a test, i want to have an algo that's guaranteed to be connected instead 2011-11-10T09:59:19 yeah 2011-11-10T09:59:20 i don't feel like generating maps until the test is positive 2011-11-10T09:59:28 i don't think i check connected anymore 2011-11-10T09:59:49 you could do something similar to my mapget 2011-11-10T09:59:54 mapgen 2011-11-10T09:59:59 delt0r: you have a mapgen too? 2011-11-10T10:00:03 I do the random walk thing 2011-11-10T10:00:15 yes, i'm looking at that too 2011-11-10T10:00:16 then i flood fill all non water regions 2011-11-10T10:00:20 *** rb_ has quit IRC (Ping timeout: 252 seconds) 2011-11-10T10:00:33 *** cbad has quit IRC (Read error: Connection reset by peer) 2011-11-10T10:00:54 keep the biggest region and convert all smaller regions to water... One connected area guaranteed 2011-11-10T10:01:03 i could never get random walks working well while starting them all from ant hills 2011-11-10T10:01:25 delt0r: do you have the code somewhere 2011-11-10T10:01:28 add hills as required 2011-11-10T10:01:38 * delt0r look around 2011-11-10T10:02:01 amstan: i have a feeling the maze algorithm gave me connectedness 2011-11-10T10:02:07 but i'm not positive 2011-11-10T10:02:18 antimatroid: i have to implement the maze too 2011-11-10T10:02:53 diagonal mazes wouldn't be bad either 2011-11-10T10:02:57 i think more awesome mazes could be done, but i found it really hard to do, especially with wider corridors 2011-11-10T10:03:08 like if you random walked a block around to make a maze 2011-11-10T10:03:15 so the corridors weren't all straight etc. 2011-11-10T10:03:18 *** bergmark has quit IRC (Ping timeout: 245 seconds) 2011-11-10T10:03:23 oh, yeah, that would be nice 2011-11-10T10:03:38 i could never get that going either 2011-11-10T10:03:39 amstan: http://pastebin.com/cHRhrGEy 2011-11-10T10:03:44 not the best code 2011-11-10T10:03:45 i couldn't* 2011-11-10T10:03:46 java? 2011-11-10T10:03:51 but i like the maps 2011-11-10T10:03:52 i could never is terrible phrasing 2011-11-10T10:03:57 that is long 2011-11-10T10:04:11 amstan: yes its java, but you get the idea... its a hacky but not much code 2011-11-10T10:04:21 no magic IIRC ;) 2011-11-10T10:04:40 here's my cavemap: https://github.com/aichallenge/aichallenge/blob/amstan_mapgen/ants/mapgen/cavemap.py 2011-11-10T10:04:40 * antimatroid looks at mine which is 641 lines of python ah 2011-11-10T10:04:46 38 lines 2011-11-10T10:04:53 without counting my other files 2011-11-10T10:05:19 *** xbelt has quit IRC (Ping timeout: 248 seconds) 2011-11-10T10:05:19 yea was about to say that 2011-11-10T10:05:34 without the rest of the code in other words... 2011-11-10T10:05:44 but like.. once you have symmetry setup, you can do map[point]=WATER 2011-11-10T10:05:54 *** cbad has joined #aichallenge 2011-11-10T10:05:54 *** cbad has joined #aichallenge 2011-11-10T10:05:55 and it'll set all points that are symmetric to point to the same 2011-11-10T10:06:03 I don't think i missing anything more than a enumeration for symmetries 2011-11-10T10:06:14 *** rb_ has joined #aichallenge 2011-11-10T10:06:15 delt0r: can you do translational symmetry? 2011-11-10T10:06:24 its all i do IIRC 2011-11-10T10:06:46 it was a first cut back when we had not even decided to do ants 2011-11-10T10:07:23 http://paste.pocoo.org/show/505802/ 2011-11-10T10:07:24 and i manage it all manually...Java with C style code :D 2011-11-10T10:07:48 do you have preprocessors in java? 2011-11-10T10:08:00 no 2011-11-10T10:08:03 aww 2011-11-10T10:08:12 good thing if you ask me 2011-11-10T10:08:29 meh.. i guess you can do objects and stuff 2011-11-10T10:08:38 so you could do everything but map[point] overloading 2011-11-10T10:08:51 C macros aka preprocessors are a mistake... put in real macros instead aka lisp 2011-11-10T10:09:08 nvm, you can't overload addition between objects 2011-11-10T10:09:15 so point1+point2 won't work to translate 2011-11-10T10:09:27 like this: https://github.com/aichallenge/aichallenge/blob/amstan_mapgen/ants/mapgen/util.py 2011-11-10T10:09:32 amstan: the java way would be a map class with map.set(int x,int y,int value) 2011-11-10T10:09:43 *** caution has quit IRC (Quit: caution) 2011-11-10T10:09:50 *** ifdef has left #aichallenge 2011-11-10T10:09:53 *** willvarfar has quit IRC (Quit: Leaving) 2011-11-10T10:09:55 or with a point class point.add(point2) 2011-11-10T10:09:59 delt0r: note how i don't have x and y in the higher level code 2011-11-10T10:10:04 but meh for this sort of thing 2011-11-10T10:10:12 i often just code it up 2011-11-10T10:10:50 for point in self.size.upto(): to iterate through all the spots on the map 2011-11-10T10:12:43 that mapgenerator took me about 1-2 hours to code up 2011-11-10T10:13:02 sure its not production qualitilty... but it was a prototype 2011-11-10T10:13:48 how do I set the name of my bot? 2011-11-10T10:14:00 Insti: it's your username 2011-11-10T10:14:10 on my local system 2011-11-10T10:14:11 MyBot. 2011-11-10T10:14:16 when I'm running play_one_game 2011-11-10T10:14:23 Why would you care about that 2011-11-10T10:14:39 it names them player_1 player_2 player_3 player_4 2011-11-10T10:14:43 On my system ith's the filename 2011-11-10T10:14:49 I'd like to know which one is mine 2011-11-10T10:14:52 Oh 2011-11-10T10:15:08 *** treeform has joined #aichallenge 2011-11-10T10:15:16 That is a bit confusing 2011-11-10T10:15:16 Insti: the order you pass them to playgame.py is the player number 2011-11-10T10:15:21 hmm it worked in the tutorial 2011-11-10T10:15:26 if you are first then your player 1 2011-11-10T10:16:41 I use the playgame.py, no idea what the difference is 2011-11-10T10:17:35 *** DeGi has quit IRC (Ping timeout: 265 seconds) 2011-11-10T10:17:59 *** Harpyon has joined #aichallenge 2011-11-10T10:18:32 *** bergmark has joined #aichallenge 2011-11-10T10:19:05 *** Khaki has joined #aichallenge 2011-11-10T10:19:07 *** xathis has joined #aichallenge 2011-11-10T10:19:16 *** delt0r_ has quit IRC (Ping timeout: 258 seconds) 2011-11-10T10:19:59 *** _ud <_ud!c3dabef0@gateway/web/freenode/ip.195.218.190.240> has quit IRC (Quit: Page closed) 2011-11-10T10:23:30 a1k0n: good game but your final run when attacking should be done with more ants, not just a 1xn row, http://aichallenge.org/visualizer.php?game=81629 2011-11-10T10:24:44 working on it 2011-11-10T10:25:06 i submitted that bot right after it got some very rudimentary combat awareness 2011-11-10T10:25:42 i just implemented an idea i really like 2011-11-10T10:27:04 Hey guys, whats the heighest number of ants you have reach in one single round? 2011-11-10T10:29:07 that map is sweet 2011-11-10T10:31:41 *** Savaron has quit IRC (Quit: Page closed) 2011-11-10T10:31:43 *** delt0r_ has joined #aichallenge 2011-11-10T10:32:15 NoxiaZ^: what do you mean ? during a single game ? 2011-11-10T10:32:26 on the official server ? 2011-11-10T10:36:08 *** jasox has joined #aichallenge 2011-11-10T10:38:44 *** olexs has quit IRC (Ping timeout: 260 seconds) 2011-11-10T10:39:49 *** JanneP__ has joined #aichallenge 2011-11-10T10:39:55 why do a lot of bots, even the top ones, jitter ant movements? i.e. repeatedly move between two squares? 2011-11-10T10:41:54 liberforce, yes on the official server. - The reason why i ask is becuase of the performce, i know how many right now i can control at same time, if its get over, it will be hard. 2011-11-10T10:41:55 *** aerique has quit IRC (Quit: ...) 2011-11-10T10:41:59 Khaki: simple rules often lead to that behavior 2011-11-10T10:42:29 even more complex ones don't rule it out 2011-11-10T10:42:38 So lets say if i can control 200 ants before the 500 ms, and you never reach more then 100 ants, i dont wanna spend more time on performce 2011-11-10T10:42:40 *** JanneP_ has quit IRC (Ping timeout: 255 seconds) 2011-11-10T10:42:49 I see 2011-11-10T10:43:06 there's actually another good reason for it 2011-11-10T10:43:14 stationary ants may prevent food from spawning until they move 2011-11-10T10:43:37 *** marijnfs has joined #aichallenge 2011-11-10T10:43:38 NoxiaZ^: there is really no upper bound to the amount of ants you can get other than than the upper bound of food for a given map 2011-11-10T10:44:19 delt0r, ahh oki 2011-11-10T10:44:19 antimatroid, Khaki: also you see more if you move... however that does not really mean moving back and forward one square 2011-11-10T10:45:00 NoxiaZ^: sometimes i get a timeout around 130, sometimes at 170 2011-11-10T10:45:19 antimatroid, in that case it seems like a quirk in the rules 2011-11-10T10:45:58 hard to keep food spawning symmetric otherwise 2011-11-10T10:46:15 liberforce, ye and thats pretty bad - Right now with this setup i will be able to handle 100 ants, but if it gets higher it will be hard 2011-11-10T10:46:23 And then i need to comeup with some new idea 2011-11-10T10:46:23 unless you allows bots to potentially affect where food can/'t spawn elsewhere on the map 2011-11-10T10:46:24 I doubt many bots are doing this from a food spawning perspective 2011-11-10T10:46:57 delt0r_: i would suggest it might be worth moving non-moved ants at the end of each turn for this purpose 2011-11-10T10:47:11 NoxiaZ^: if you're afraid to timeout if you have to many ants, just make an ant remain on top of your hill once you've reach the count 2011-11-10T10:47:17 so if you want an ant to be stationary make that as a "move" with your bot 2011-11-10T10:47:37 oh, it's 11/11/11 2011-11-10T10:47:53 antimatroid: not here ;) 2011-11-10T10:48:09 liberforce: i'm from the future :) 2011-11-10T10:48:25 wonders of IPoT :) 2011-11-10T10:50:13 liberforce, ahh that was smart :D 2011-11-10T10:50:47 I would rather find the timehog and fix it 2011-11-10T10:51:16 pairofdice, there will be a point where you cant do it between the 500 ms 2011-11-10T10:51:30 NoxiaZ^: that way if you fall below your max ant count, you can make that one move and another will spawn 2011-11-10T10:51:35 thats just impossible to fix 2011-11-10T10:51:48 liberforce, thats really good idea! :) 2011-11-10T10:51:54 The number of ants I have don't considerably raise my turntimes 2011-11-10T10:52:42 *** jasox has quit IRC (Quit: Leaving) 2011-11-10T10:53:02 Well, it does of course, but I don't do heavy lifting for each indicidual ant 2011-11-10T10:53:17 A bit more in combat 2011-11-10T10:53:30 *** amstan has quit IRC (Ping timeout: 276 seconds) 2011-11-10T10:53:34 NoxiaZ^: moreover and as long you have an ant on the hill, and ants in the hive, your hill can't be destroyed 2011-11-10T10:53:35 So you just said it you self 2011-11-10T10:53:57 liberforce, ahh i didnt know that - thanks :) 2011-11-10T10:53:58 I do 300 ants just fine 2011-11-10T10:54:23 pairofdice, so again if its possible to get 400 ants you wont be able to handle all 2011-11-10T10:54:37 pairofdice: do you use an anti-object approach ? 2011-11-10T10:54:39 And as i dont know how many ants you can get in a single round, you have to deal with that problem 2011-11-10T10:54:47 I haven't gotten to 400 ants yet 2011-11-10T10:55:15 *** lorill has joined #aichallenge 2011-11-10T10:55:29 pairofdice: 300 ants out, but how many do you move each turn ? all of them ? 2011-11-10T10:55:37 All 2011-11-10T10:56:03 *** hanshoi has joined #aichallenge 2011-11-10T10:56:03 i've noticed that sometimes not moving is the best local move 2011-11-10T10:57:16 NoxiaZ^, yeah, but I would still rather make my inefficient code more efficient than block ants from spawning 2011-11-10T10:57:29 If I can't do that anymore, then sure, block ants 2011-11-10T10:57:59 ofc i will also do my code more effictive before blocking 2011-11-10T10:58:11 i havnt reach that point yet. i just have to think about everything from start 2011-11-10T10:58:20 rather then doing it at the end 2011-11-10T10:58:37 pairofdice: your bot is a python3 one, right ? 2011-11-10T10:59:43 *** lorill has quit IRC (Client Quit) 2011-11-10T11:00:11 Yeah 2011-11-10T11:00:55 *** kilae has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111104165243]) 2011-11-10T11:08:20 *** JorgeB has quit IRC (Quit: Computer has gone to sleep.) 2011-11-10T11:13:19 *** bergmark has quit IRC (Ping timeout: 258 seconds) 2011-11-10T11:20:39 *** superice has joined #aichallenge 2011-11-10T11:20:52 hi all 2011-11-10T11:24:32 *** JorgeB has joined #aichallenge 2011-11-10T11:25:48 *** bergmark has joined #aichallenge 2011-11-10T11:26:00 Oh man, just when I finished saying my bot doesn't time out 2011-11-10T11:26:12 In addition to that it gained some rather ill behaviour 2011-11-10T11:28:19 pairofdice: i'ts already amazing tat you can control that many ants with a python 3 bot. Python has a little sluggish reputation 2011-11-10T11:28:40 pairofdice: do you use anti-object as main orientation 2011-11-10T11:28:40 ? 2011-11-10T11:29:15 I don't think so 2011-11-10T11:29:53 well do you directly control the ants, or do you perform the computation on the tiles and let the ants go their own way ? 2011-11-10T11:30:26 I guess I do in that I calculate one map and let ants decide where they go 2011-11-10T11:30:47 ok, so it's an anti-object approach 2011-11-10T11:30:55 pairofdice: thanks for the info 2011-11-10T11:31:07 I was already on the way to convert my bot to that 2011-11-10T11:31:29 guiding the ants is too complex and CPU-hungry 2011-11-10T11:31:34 Well I'm not someone you want to take programming advice from :) 2011-11-10T11:31:43 pairofdice: why ? 2011-11-10T11:32:17 your revision 14 had a mu = 70 which is way higher than mine ;) 2011-11-10T11:32:30 *** Extrarius has quit IRC (Ping timeout: 240 seconds) 2011-11-10T11:32:49 i never went over 56 2011-11-10T11:32:55 My coding experience sums up to about a month 2011-11-10T11:33:03 :'( 2011-11-10T11:33:13 mine is around 10 years 2011-11-10T11:35:22 Maybe a month is a bit of anti-exaggeration 2011-11-10T11:35:23 *** foRei has joined #aichallenge 2011-11-10T11:35:46 How do you let your ants decide where to go? 2011-11-10T11:36:12 I make a mountain, they climb it 2011-11-10T11:36:33 What do they decide out from? - And you dont use the nearest and to pick up food? 2011-11-10T11:36:50 NoxiaZ^: http://www.cs.colorado.edu/~ralex/papers/PDF/OOPSLA06antiobjects.pdf 2011-11-10T11:38:52 NoxiaZ^, nah, figuring out which one is closest it rather expensive 2011-11-10T11:39:06 is 2011-11-10T11:39:46 http://queue.acm.org/icpc/ 2011-11-10T11:39:50 anyone looked at that? 2011-11-10T11:41:24 pairofdice, and thats the reason why i wanna know how many ants you will get in a single round 2011-11-10T11:42:08 i can control around 100 ants by checking the distance to every obejct and allways select the closest 2011-11-10T11:42:22 antimatroid: nice, tnaks 2011-11-10T11:42:26 thanks 2011-11-10T11:42:37 NoxiaZ^, yeah, that scales very badly 2011-11-10T11:43:02 Yes, but if i dont reach 100 ants its not a problem 2011-11-10T11:43:50 NoxiaZ^: the turntime variable is set to 500ms now 2011-11-10T11:44:00 yes 2011-11-10T11:44:00 *** links234 has joined #aichallenge 2011-11-10T11:44:02 and thats enough 2011-11-10T11:44:08 right now 2011-11-10T11:44:11 NoxiaZ^: nothing tells that they won't test with another value during the final round 2011-11-10T11:44:25 which makes your 100 ants constant pointless 2011-11-10T11:44:49 True if they change it it would 2011-11-10T11:45:00 but i think they will 2011-11-10T11:45:04 even that it is possible 2011-11-10T11:45:13 *** JorgeB has quit IRC (Ping timeout: 258 seconds) 2011-11-10T11:45:54 It makes no sense to make something 500 ms, and you make you code out from that, and when everyone is done then change it. 2011-11-10T11:46:08 hmm... the bots run as independent processes that communicate over I/O? 2011-11-10T11:46:39 NoxiaZ^: it's called adaptability 2011-11-10T11:46:43 or AI 2011-11-10T11:47:03 yes p_l, I/O being stdin and stdout 2011-11-10T11:47:28 ... sweet 2011-11-10T11:47:46 *** JorgeB has joined #aichallenge 2011-11-10T11:47:54 But still it wont be that big a problem, because the time is track able anyways 2011-11-10T11:48:08 p_l: a bot being all ants from one player, in case that was unclear :) 2011-11-10T11:48:16 mleise: yes, I know :) 2011-11-10T11:48:29 does anyone here tracks time to avoid timeout ? 2011-11-10T11:48:36 But when talking about AI liber, there is no AI in letting the walk "randomly" or not use the closest one 2011-11-10T11:48:50 most of the starting kits track time 2011-11-10T11:48:53 NoxiaZ^: ... actually there is 2011-11-10T11:49:18 so just check the remaining time at a sensible place in your code 2011-11-10T11:49:25 AI covers surprisingly a lot, except that as soon as something works and it isn't a game, it's not called AI 2011-11-10T11:49:36 :) 2011-11-10T11:50:27 p_l, i dont get you :) 2011-11-10T11:50:33 *** speedway has joined #aichallenge 2011-11-10T11:50:38 *** TheLinker has joined #aichallenge 2011-11-10T11:50:40 which is why after AI Winter we invented bullshit buzzwords like "machine learning", "knowledge-based engineering", etc ;-) 2011-11-10T11:51:10 Minthos: the starter bot in C din't track time ;) 2011-11-10T11:51:41 there is no intelligence in just moving random around :) 2011-11-10T11:52:00 NoxiaZ^: it can be part of reaching the black triangle 2011-11-10T11:52:01 NoxiaZ^: not if it's completely random 2011-11-10T11:52:30 but if you want to avoid moving an ant where another already is, or avoid water, that's intelligence 2011-11-10T11:53:06 Ye, but thats not even close enough to be effective enough :) 2011-11-10T11:53:19 a random movement generator can be useful as beginning of pipeline (with later parts doing avoidance mechanisms, or biasing, etc.) 2011-11-10T11:55:53 personally I'm going to use an influence map to guide the default movement pattern, and possibly a random wanderer or patrolling sequence for food gathering 2011-11-10T11:57:00 p_l: you won't change the map for food gathering ? 2011-11-10T11:57:37 liberforce: influence map will be for combat and exploration, mainly 2011-11-10T11:57:38 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-11-10T11:58:34 python n00b: I'm having trouble writing out debug code. any clues? sys.stderr.write('hello world') in my setup seems to not be doing anything. 2011-11-10T11:59:33 Insti: you need to add the -e option to the scripts calling playgame.py , otherwise stderr won't show 2011-11-10T11:59:44 my first turn take 23ms 2011-11-10T11:59:47 liberforce: I'm going to experiment with how orders are handled, though 2011-11-10T11:59:57 my second takes 1ms 2011-11-10T12:00:00 liberforce done that. :( 2011-11-10T12:00:08 that is the jvm warmup for you 2011-11-10T12:00:34 Insti: or -E to write stderr to a file if I recall correctly 2011-11-10T12:00:43 I'm getting 0.botN.error log files but they are empty. 2011-11-10T12:00:54 delt0r: I'm avoiding that by using a language that compiles directly :D 2011-11-10T12:01:34 meh... i like my GC... not really java per say... but since i use all the time at work... its what i can use fast these days 2011-11-10T12:01:42 going back to C/C++ 2011-11-10T12:01:49 * delt0r shivers 2011-11-10T12:02:12 p_l: what are you using? 2011-11-10T12:02:14 delt0r: I chose C from the start ;) 2011-11-10T12:02:33 easier to optimize for me 2011-11-10T12:02:44 that is an important point 2011-11-10T12:02:45 but lack of queue, lists, etc is boring 2011-11-10T12:02:59 had to reimplement some basic stuff 2011-11-10T12:03:00 what you know can make more difference to performance that the "lanuage" 2011-11-10T12:03:11 yep 2011-11-10T12:03:21 *** Giuliano has joined #aichallenge 2011-11-10T12:03:34 half the C vers java performance wars are because a C programmer writes crap java, and a java programmer bad C 2011-11-10T12:03:39 delt0r: Common Lisp 2011-11-10T12:03:41 heard python bot would easily timeout 2011-11-10T12:03:47 p_l: i like lisp 2011-11-10T12:03:57 almost used it this time round 2011-11-10T12:04:03 *** rajanaresh has joined #aichallenge 2011-11-10T12:04:15 was thinking of one of the lisp dialects on the JVM 2011-11-10T12:04:21 I get too annoyed by Java to work with it 2011-11-10T12:04:34 p_l: you do get over it... trust me 2011-11-10T12:04:35 Since my preferred implementation is available, why not use it :) 2011-11-10T12:04:40 indeed 2011-11-10T12:04:40 delt0r: no, it only grows 2011-11-10T12:05:05 My first recursion was with lisp 2011-11-10T12:05:15 back when i was a wee lad 2011-11-10T12:05:22 delt0r: Java the language itself, or JVM as the bytecode platform, isn't that bad... but then the libs come in... *pukes* 2011-11-10T12:05:34 its the libs i like 2011-11-10T12:05:53 delt0r: I dislike certain patterns that are very popular in many libs in Java. 2011-11-10T12:06:01 I'm going to stay with Common Lisp :) 2011-11-10T12:06:08 seriously diffenct tcp code on different OSes but same hardware in 2011? WTF? 2011-11-10T12:06:41 p_l: that is the same in every land IMO ...and even experience. There is always some clug 2011-11-10T12:06:48 lang 2011-11-10T12:06:51 not land 2011-11-10T12:07:11 but there is a government analogy in there somewhere ;) 2011-11-10T12:07:20 p_l: how is your bot going? 2011-11-10T12:07:29 delt0r: the difference for me is that CL gives me enough power to technically rewrite anything I need fast... which is annoying when you're getting detracted from writing something by urge to implement your own database 2011-11-10T12:07:40 lol 2011-11-10T12:07:49 *** peyton has quit IRC (Quit: peyton) 2011-11-10T12:07:51 delt0r: barely started - I'm swamped with other (home)work 2011-11-10T12:08:04 well i have that problem in java too... but yea, for example the CLOS is great 2011-11-10T12:08:42 why doesn't java have double dispatch! 2011-11-10T12:11:40 *** speedway has quit IRC (Ping timeout: 258 seconds) 2011-11-10T12:12:10 *** mj41 has quit IRC (Read error: Operation timed out) 2011-11-10T12:12:56 haha cl has multiple 2011-11-10T12:13:24 honestly, i thought it would be harder to wirte my bot in c++ 2011-11-10T12:13:46 i really miss C-c C-c though 2011-11-10T12:14:00 its a bitch that i need to recompile to see a change 2011-11-10T12:14:16 and those sweet run-time function redefinitions! 2011-11-10T12:15:34 *** ztfw has joined #aichallenge 2011-11-10T12:16:20 :) 2011-11-10T12:17:03 mcstar: don't forget DISASSEMBLE and runtime access to assembler and compiler :D 2011-11-10T12:17:15 (at least in SBCL, Allegro and CCL) 2011-11-10T12:17:27 i would lie if i said i used those 2011-11-10T12:17:28 *** rajanaresh has quit IRC (Ping timeout: 256 seconds) 2011-11-10T12:17:36 i was a newbie 2011-11-10T12:17:52 p_l: what i missed the most in cl, interestingly is the compiler 2011-11-10T12:17:58 I know a little of how to use SBCL's assembler 2011-11-10T12:18:10 meaning, its easier for me to write correct c++ code, beacuse it catches type error 2011-11-10T12:18:12 s 2011-11-10T12:18:36 mcstar: use SBCL with full safety on. You'll get tons of warnings :) 2011-11-10T12:18:48 ill do next time 2011-11-10T12:19:13 also, you can catch more by defining the types 2011-11-10T12:19:16 i decided to learn c++ too, cause itnerestingly there are jobs offers requring c++ skill 2011-11-10T12:19:20 *** speedway has joined #aichallenge 2011-11-10T12:19:36 though if you really want to "write correct code", go for Haskell 2011-11-10T12:19:45 p_l: i only defined types when i wante to optimize a debugged code 2011-11-10T12:20:03 mcstar: at full safety, the types will often work as assertion 2011-11-10T12:20:05 maybe that was a mistake 2011-11-10T12:20:14 *** amstan has joined #aichallenge 2011-11-10T12:20:14 *** ChanServ sets mode: +o amstan 2011-11-10T12:20:20 well, it depends on your programming style 2011-11-10T12:20:34 you can as well just put assertions where you want 2011-11-10T12:20:45 *** xathis has quit IRC () 2011-11-10T12:23:13 *** amstan_ has joined #aichallenge 2011-11-10T12:23:13 *** ChanServ sets mode: +o amstan_ 2011-11-10T12:23:20 wow 2011-11-10T12:23:34 just saw a guy controling 514 ants with no timeout 2011-11-10T12:23:54 *** amstan has quit IRC (Read error: No route to host) 2011-11-10T12:23:57 interesting approach :) 2011-11-10T12:24:22 *** praveen_ has joined #aichallenge 2011-11-10T12:25:23 hey, got myself 299 ants before timeout 2011-11-10T12:25:45 liberforce: i am not even using 10ms with 1000 ants... 2011-11-10T12:26:06 of course they don't even know what a ant hill is ;) 2011-11-10T12:26:22 or why these other ants keep killing them! 2011-11-10T12:27:13 liberforce my problem was not flushing 2011-11-10T12:27:21 delt0r: so they know they are dead? whoa 2011-11-10T12:27:34 liberforce: are you saving paths between turns or are they all recalculating? 2011-11-10T12:28:19 *** Antimony has joined #aichallenge 2011-11-10T12:29:31 *** Blkt has quit IRC (Quit: ERC Version 5.3 (IRC client for Emacs)) 2011-11-10T12:30:50 *** ikaros has joined #aichallenge 2011-11-10T12:33:15 *** speedway has quit IRC (Read error: Connection reset by peer) 2011-11-10T12:33:28 *** aarossig has quit IRC (Remote host closed the connection) 2011-11-10T12:33:56 delt0r: which language are you using ? 2011-11-10T12:34:02 *** speedway has joined #aichallenge 2011-11-10T12:34:19 twymer: I recalculate 2011-11-10T12:34:22 *** Giuliano has quit IRC (Ping timeout: 265 seconds) 2011-11-10T12:36:18 Is there any way to tell if a hill has been destroyed other than if you razed it yourself? In other words just because you don't see a hill there doesn't mean its not still there but below another ant right? 2011-11-10T12:36:33 Maybe as an exercise in futility I try implementing my bot in c++ 2011-11-10T12:40:28 *** speedway has quit IRC (Read error: Connection reset by peer) 2011-11-10T12:40:40 bergmark: I believe the hills positions are told, regardles of the fact that there is a ant on it or not. 2011-11-10T12:41:33 it's after that up to the starter package how to use that information. 2011-11-10T12:42:12 bergmark: the engine only tells you there's a hill if it wasn't rased already 2011-11-10T12:42:32 so if you don't recieve info about a hill that you remember being there it means it was razed 2011-11-10T12:42:45 hanshoi: I see, thank you. 2011-11-10T12:43:19 amstan_: Only if its in visible range right? 2011-11-10T12:43:25 yes 2011-11-10T12:47:53 liberforce: java 2011-11-10T12:48:32 *** superice has quit IRC (Quit: Page closed) 2011-11-10T12:51:18 *** xbelt has joined #aichallenge 2011-11-10T12:52:19 *** Fandekasp has joined #aichallenge 2011-11-10T12:52:40 *** dorisabayon has joined #aichallenge 2011-11-10T12:53:57 delt0r: ok, so your ants are pretty dumb, but they wanna win by the number... 2011-11-10T12:54:09 lol 2011-11-10T12:54:20 *** xbelt has quit IRC (Remote host closed the connection) 2011-11-10T12:54:27 *** xbelt has joined #aichallenge 2011-11-10T12:54:45 *** treeform has quit IRC (Remote host closed the connection) 2011-11-10T12:55:11 either pretty, or dumb, you cant have both 2011-11-10T12:56:51 k, see you guys ! 2011-11-10T12:57:02 *** liberforce has left #aichallenge 2011-11-10T12:57:25 *** xbelt has quit IRC (Remote host closed the connection) 2011-11-10T12:57:33 *** xbelt has joined #aichallenge 2011-11-10T12:57:41 my ants are not dumb... the just lack perception of combat or hills 2011-11-10T13:00:35 *** xbelt has quit IRC (Remote host closed the connection) 2011-11-10T13:00:42 *** xbelt has joined #aichallenge 2011-11-10T13:01:24 *** dvladim has joined #aichallenge 2011-11-10T13:04:05 as for running around getting all the food they can... they are really good at it 2011-11-10T13:04:45 *** amstan_ has quit IRC (Ping timeout: 260 seconds) 2011-11-10T13:04:55 *** epicmonkey has joined #aichallenge 2011-11-10T13:07:41 *** xbelt has quit IRC (Remote host closed the connection) 2011-11-10T13:07:49 *** xbelt has joined #aichallenge 2011-11-10T13:09:42 my bot's the same 2011-11-10T13:09:56 except it also tries to guard the hills 2011-11-10T13:10:00 hahah 2011-11-10T13:10:03 ive got a great idea 2011-11-10T13:10:04 but it isn't very efficient due to lack of combat skill 2011-11-10T13:10:34 dont any one of you want to convince my friend to enter the competition? 2011-11-10T13:10:39 i would appreciate it 2011-11-10T13:10:52 *** xbelt has quit IRC (Remote host closed the connection) 2011-11-10T13:10:55 just type /query zoli1 2011-11-10T13:11:00 so? 2011-11-10T13:11:18 *** xbelt has joined #aichallenge 2011-11-10T13:11:28 he is a bit afraid, new to programming, but he is a smart guy, so he can handle this competition 2011-11-10T13:12:59 try not to spam people 2011-11-10T13:14:01 *** speedway has joined #aichallenge 2011-11-10T13:14:16 *** xbelt has quit IRC (Remote host closed the connection) 2011-11-10T13:14:24 *** xbelt has joined #aichallenge 2011-11-10T13:15:26 One problem solved, ten other problems crop up 2011-11-10T13:19:34 Zannick: ? 2011-11-10T13:19:59 mcstar: you just asked a channel of 170+ people to message someone 2011-11-10T13:20:11 and probably noone will 2011-11-10T13:20:21 several thousand would be ideal 2011-11-10T13:21:40 *** xbelt has quit IRC (Read error: Connection reset by peer) 2011-11-10T13:23:27 *** Cyndre has joined #aichallenge 2011-11-10T13:25:04 *** Antimony_ has joined #aichallenge 2011-11-10T13:25:46 *** xbelt has joined #aichallenge 2011-11-10T13:27:30 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-10T13:28:29 *** mviel_ has quit IRC (Remote host closed the connection) 2011-11-10T13:30:31 *** Antimony_ has quit IRC (Ping timeout: 255 seconds) 2011-11-10T13:30:37 double free or corruption (out): 2011-11-10T13:30:37 *** xbelt has quit IRC (Remote host closed the connection) 2011-11-10T13:30:47 *** xbelt has joined #aichallenge 2011-11-10T13:32:21 weird 2011-11-10T13:32:52 *** grwip has quit IRC (Remote host closed the connection) 2011-11-10T13:32:57 ive got to take divisablity into accoutn 2011-11-10T13:33:03 *** mrNub has quit IRC (Quit: Page closed) 2011-11-10T13:33:16 *** peyton has joined #aichallenge 2011-11-10T13:34:01 *** datachomper has joined #aichallenge 2011-11-10T13:34:58 *** peyton has quit IRC (Client Quit) 2011-11-10T13:35:37 *** xbelt has quit IRC (Remote host closed the connection) 2011-11-10T13:35:44 *** xbelt has joined #aichallenge 2011-11-10T13:38:54 *** praveen_ has quit IRC (Quit: Page closed) 2011-11-10T13:41:12 *** the-mgt has quit IRC (Quit: the-mgt) 2011-11-10T13:41:21 *** xbelt has quit IRC (Remote host closed the connection) 2011-11-10T13:41:32 *** xbelt has joined #aichallenge 2011-11-10T13:42:13 *** speedway has quit IRC (Ping timeout: 255 seconds) 2011-11-10T13:45:21 *** xbelt has quit IRC (Remote host closed the connection) 2011-11-10T13:45:28 *** xbelt has joined #aichallenge 2011-11-10T13:46:23 *** Conorach has joined #aichallenge 2011-11-10T13:46:30 *** Fandekasp has quit IRC (Ping timeout: 240 seconds) 2011-11-10T13:46:39 *** dorisabayon has quit IRC (Ping timeout: 248 seconds) 2011-11-10T13:47:18 *** treeform has joined #aichallenge 2011-11-10T13:48:26 *** xbelt has quit IRC (Remote host closed the connection) 2011-11-10T13:48:35 *** xbelt has joined #aichallenge 2011-11-10T13:49:11 aha 2011-11-10T13:49:21 so for example vector vec; 2011-11-10T13:49:29 doesnt initialize the vector? 2011-11-10T13:49:41 i cant call .size() on it? 2011-11-10T13:50:17 vector vec(size, default); 2011-11-10T13:50:56 normally done in constructor initialization. 2011-11-10T13:51:26 no, its ook 2011-11-10T13:51:37 ok, sorry :) 2011-11-10T13:51:43 you see, i could always use a vector if i did like that 2011-11-10T13:51:56 but now i got a segfault, but it seems because of something else 2011-11-10T13:52:10 :( 2011-11-10T13:52:15 i could list all the sizes of my vectors, so they must have got initialized 2011-11-10T13:54:32 valgrind says im using something uninitialized 2011-11-10T13:54:49 *** slayer has joined #aichallenge 2011-11-10T13:56:39 mcstar: Are you still having trouble with stderr out? 2011-11-10T13:56:51 i never had 2011-11-10T13:56:58 turned out i was just dumb 2011-11-10T13:57:08 What was your problem? I'm not seeing anything myself ... 2011-11-10T13:57:26 my problem was, that i searched for "turn" 2011-11-10T13:57:28 *** mleise has quit IRC (Quit: Leaving.) 2011-11-10T13:57:38 and i saw turn 109 bot 3 crashed 2011-11-10T13:57:40 or similar 2011-11-10T13:57:49 but i only recognized "turn 109" 2011-11-10T13:58:01 and i wondered where are my other 108 lines from the log 2011-11-10T13:58:38 *** xbelt has quit IRC (Read error: Connection reset by peer) 2011-11-10T13:58:48 why just write to a file instead of stderr? 2011-11-10T13:58:57 thats what i did and have had no problems at all 2011-11-10T13:59:24 *** xbelt has joined #aichallenge 2011-11-10T14:00:22 treeform: the engine does that for you, so why bother? 2011-11-10T14:00:45 oh i did not know that 2011-11-10T14:01:02 it's supposed to, anyway 2011-11-10T14:01:17 i though you would have to capture 2011-11-10T14:01:48 -e or -E does that 2011-11-10T14:02:00 and the --log_dir= option 2011-11-10T14:02:05 or similar 2011-11-10T14:02:11 ./playgame tells you all 2011-11-10T14:02:17 well i guess i solved it myself, without knowing 2011-11-10T14:03:37 *** Antimony has joined #aichallenge 2011-11-10T14:03:59 *** xbelt has quit IRC (Read error: Connection reset by peer) 2011-11-10T14:04:16 *** praveen has quit IRC (Ping timeout: 265 seconds) 2011-11-10T14:06:13 *** xbelt has joined #aichallenge 2011-11-10T14:06:24 mcstar: this is what my bot generates as it runs (wait till it loads then use arrow keys) http://pastehtml.com/view/bdkbos3wy.html 2011-11-10T14:06:59 *** the-mgt has joined #aichallenge 2011-11-10T14:08:19 treeform: nice, but dont compress the images 2011-11-10T14:08:32 ok, i see 2011-11-10T14:08:39 than just disable the linear filtering 2011-11-10T14:08:51 *** xbelt has quit IRC (Read error: Operation timed out) 2011-11-10T14:08:55 or you dont like it "pixelated"? 2011-11-10T14:09:20 *** xbelt has joined #aichallenge 2011-11-10T14:09:49 treeform: i wish i could share with you what im doing now 2011-11-10T14:10:00 but you know, im gonna win the contest because of it 2011-11-10T14:10:13 treeform: that's cool 2011-11-10T14:10:37 treeform: yep, actually i think others could benefit from it 2011-11-10T14:10:44 what language? 2011-11-10T14:11:10 ah isee 2011-11-10T14:11:28 you have many rendering, and you overlay them on top of each other 2011-11-10T14:12:38 *** lavalamp has joined #aichallenge 2011-11-10T14:13:53 *** peyton has joined #aichallenge 2011-11-10T14:14:24 *** Antimony has quit IRC (Ping timeout: 252 seconds) 2011-11-10T14:16:55 *** slayer has quit IRC (Quit: Page closed) 2011-11-10T14:18:18 *** xbelt has quit IRC (Remote host closed the connection) 2011-11-10T14:18:25 *** xbelt has joined #aichallenge 2011-11-10T14:19:06 *** ReW has joined #aichallenge 2011-11-10T14:19:11 hello 2011-11-10T14:20:22 some1? 2011-11-10T14:20:43 433110 2011-11-10T14:21:04 oops, one 3 too much 2011-11-10T14:21:23 *** xbelt has quit IRC (Remote host closed the connection) 2011-11-10T14:21:30 *** xbelt has joined #aichallenge 2011-11-10T14:21:33 *** Khaki has quit IRC (Quit: Leaving) 2011-11-10T14:22:15 can som1 help me? 2011-11-10T14:22:51 ReW: Might be better to just ask your question. 2011-11-10T14:23:12 *** peyton has quit IRC (Quit: peyton) 2011-11-10T14:26:29 *** treeform has quit IRC (Remote host closed the connection) 2011-11-10T14:27:01 *** treeform has joined #aichallenge 2011-11-10T14:27:08 i was trying to start with this and i was following the tutorial but im already stuck i dont know where to put this: private Map orders = new HashMap(); 2011-11-10T14:27:13 *** xbelt has quit IRC (Read error: Connection reset by peer) 2011-11-10T14:27:48 *** xbelt has joined #aichallenge 2011-11-10T14:30:26 *** ReW has quit IRC () 2011-11-10T14:30:28 .. and the problem is ? (ReW) 2011-11-10T14:30:34 just missed him 2011-11-10T19:55:44 *** contestbot has joined #aichallenge 2011-11-10T19:56:35 *** Fandekasp has quit IRC (Ping timeout: 260 seconds) 2011-11-10T19:57:15 *** Fandekasp has joined #aichallenge 2011-11-10T20:00:52 *** dorisabayon has quit IRC (Ping timeout: 258 seconds) 2011-11-10T20:01:40 *** dorisabayon has joined #aichallenge 2011-11-10T20:02:00 *** jasox has quit IRC (Remote host closed the connection) 2011-11-10T20:03:15 *** Fandekasp has quit IRC (Ping timeout: 256 seconds) 2011-11-10T20:08:46 *** Fandekasp has joined #aichallenge 2011-11-10T20:10:21 *** replore has joined #aichallenge 2011-11-10T20:15:03 *** Fandekasp has quit IRC (Ping timeout: 258 seconds) 2011-11-10T20:15:06 The countdown to next game estimate must be the world's most unreliable clock 2011-11-10T20:15:14 it's worse than the old windows file copy progress bar 2011-11-10T20:16:04 *** Fandekasp has joined #aichallenge 2011-11-10T20:17:54 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-10T20:21:07 *** b0rder_ has quit IRC (Quit: 离开) 2011-11-10T20:21:16 *** b0rder_ has joined #aichallenge 2011-11-10T20:22:53 *** dorisabayon has quit IRC (Ping timeout: 276 seconds) 2011-11-10T20:23:35 *** dorisabayon has joined #aichallenge 2011-11-10T20:24:51 *** amstan has joined #aichallenge 2011-11-10T20:24:51 *** ChanServ sets mode: +o amstan 2011-11-10T20:25:42 BenJackson: I like when it shows a negative time :) 2011-11-10T20:25:52 @rankings 2011-11-10T20:25:53 amstan: Top 10 players: xathis(90.5), GreenTea(84.5), MomoBot(84.5), pguillory(84.0), GarySWest(83.9), ThisIsNotABug(83.7), sir_macelon(83.5), Migi32(82.1), lordap(81.5), Fluxid(81.4) 2011-11-10T20:26:50 oh, my bot is top 10? Awesome :) 2011-11-10T20:26:51 *** Fandekasp has quit IRC (Ping timeout: 258 seconds) 2011-11-10T20:27:00 it's been #11 for ages :p 2011-11-10T20:27:37 *** Fandekasp has joined #aichallenge 2011-11-10T20:29:58 *** dorisabayon has quit IRC (Ping timeout: 245 seconds) 2011-11-10T20:30:39 *** roflmao has joined #aichallenge 2011-11-10T20:31:16 *** dorisabayon has joined #aichallenge 2011-11-10T20:32:09 *** Antimony has joined #aichallenge 2011-11-10T20:32:24 *** amstan_ has joined #aichallenge 2011-11-10T20:32:24 *** ChanServ sets mode: +o amstan_ 2011-11-10T20:33:20 *** amstan has quit IRC (Ping timeout: 260 seconds) 2011-11-10T20:33:58 *** Fandekasp has quit IRC (Ping timeout: 252 seconds) 2011-11-10T20:34:55 *** Fandekasp has joined #aichallenge 2011-11-10T20:46:45 Migi32: wow good job 2011-11-10T20:48:51 *** amstan_ is now known as amstan 2011-11-10T20:49:31 it's not that hard really. All you have to do is upload a decent bot and don't reupload in a long time, so the sigma of your ranking goes down. 2011-11-10T20:50:47 *** treeform has quit IRC (Remote host closed the connection) 2011-11-10T20:50:48 just look at all the bots in the top 10: almost all of them are over a week old 2011-11-10T20:58:47 *** dorisabayon has quit IRC (Ping timeout: 256 seconds) 2011-11-10T20:59:16 2nd place 4 days old 2011-11-10T20:59:25 well 5 2011-11-10T20:59:33 janzert: hey 2011-11-10T20:59:35 *** dorisabayon has joined #aichallenge 2011-11-10T20:59:52 4th place 5 days as well 2011-11-10T20:59:53 hey 2011-11-10T20:59:56 janzert: i got to tell jeff to come here sometime, apparently he's very busy 2011-11-10T21:00:16 janzert: i also broke the workers 2 days ago, then fixed them 2011-11-10T21:00:24 :) ok 2011-11-10T21:00:31 i restarted memcached, so the latest_game was gone 2011-11-10T21:00:33 then i waited 2011-11-10T21:00:37 and after an hour it didn't come back 2011-11-10T21:00:44 heh 2011-11-10T21:01:00 so i looked around, accessed a random page on the site, game_list_json i think 2011-11-10T21:01:05 and that made mysql hang 2011-11-10T21:01:07 so the workers died 2011-11-10T21:01:17 hmm 2011-11-10T21:01:31 yeah, it doesn't seem like a good idea to have a random page that can hang the whole contest 2011-11-10T21:02:03 I assume it's a query that is just taking forever long? 2011-11-10T21:02:11 probably 2011-11-10T21:02:36 janzert: also.. are we interested in game servers? 3 nodes, in total: 36 cores(3.0GHz Xeon), 430GB ram 2011-11-10T21:03:18 that's equivalent to 18 workers 2011-11-10T21:03:22 ec2 workers 2011-11-10T21:03:55 in some ways, but on the other hand I expect it to take a fair bit of work to get them running 2011-11-10T21:04:07 why? 2011-11-10T21:04:34 they really have 143GB of ram per machine? 2011-11-10T21:04:44 janzert: apparently 2011-11-10T21:05:41 janzert: problem is that we'll only be able to get them at the end of november if we do 2011-11-10T21:05:49 we haven't run more than one worker on a machine in quite a while so I would expect to run into various problems there. they are inevitably going to be different than the ec2 setup and of course we can't just kill one and get another one back with a bare OS install 2011-11-10T21:06:13 *** dorisabayon has quit IRC (Ping timeout: 245 seconds) 2011-11-10T21:06:23 they're going to have esxi installed on them, so snapshots should be fairly easy to do 2011-11-10T21:06:24 that's one of the really nice things about ec2 of course 2011-11-10T21:06:54 and anyone with access will need to be withdrawn from the contest of course 2011-11-10T21:07:23 *** dorisabayon has joined #aichallenge 2011-11-10T21:07:39 janzert: so what's the final decision? 2011-11-10T21:08:06 I don't really know at this point 2011-11-10T21:08:19 just it's not a really simple thing to add them 2011-11-10T21:08:20 ok, well, we have time still 2011-11-10T21:08:28 and we need to have to find out if we do get them 2011-11-10T21:08:38 the deal is to get them advertisment like google 2011-11-10T21:09:07 If we can get other sponsors we should create a section at the bottom of the page with the names and logos 2011-11-10T21:09:23 yeah 2011-11-10T21:09:25 and I think we could get more if we can get Jeff back around 2011-11-10T21:09:53 I wouldn't want to add more names into the header btw 2011-11-10T21:10:03 *** t2027 has joined #aichallenge 2011-11-10T21:10:06 is it possible to somehow turn of the timeout option during the tests? 2011-11-10T21:10:17 <_flag> Interesting way to take 7th place: http://aichallenge.org/visualizer.php?game=83510&user=31 :P 2011-11-10T21:10:30 t2027: I don't believe so but you can set it for really really long 2011-11-10T21:10:53 e.g. 9999999 2011-11-10T21:11:18 _flag: can you look at the pull requests regarding the C starter 2011-11-10T21:11:34 tats enough how to do it? 2011-11-10T21:11:38 thats* 2011-11-10T21:11:45 <_flag> janzert: Sure 2011-11-10T21:12:14 t2027: there's a playgame.py option --turntime I believe it is 2011-11-10T21:12:37 thanks a lot :) 2011-11-10T21:12:56 <_flag> The first one about the wrapping miscalculation is correct, I forgot to add some brackets in there 2011-11-10T21:13:50 you have commit access right? if so just click the green merge the pull request button 2011-11-10T21:14:28 aichallenge: flagcapper epsilon * r891900d / ants/dist/starter_bots/c/YourCode.c : 2011-11-10T21:14:28 aichallenge: Merge pull request #363 from nikai3d/nikai 2011-11-10T21:14:28 aichallenge: C starter pack: bug in map wrap calculation - http://git.io/NLwM_Q 2011-11-10T21:14:37 except he's changing a - to a +? 2011-11-10T21:15:27 <_flag> Yes, but had I put brackets around (Info->cols - 1) like I did in my bot it would come out to the same thing 2011-11-10T21:15:30 heh, wait 3 hours, play two games within 8 minutes 2011-11-10T21:15:39 *** dorisabayon has quit IRC (Ping timeout: 260 seconds) 2011-11-10T21:15:48 ahh, yes 2011-11-10T21:16:35 It's amazing how luck-based some games are despite all attempts to equalize things 2011-11-10T21:18:52 Whoa. Colored text. 2011-11-10T21:21:20 *** dorisabayon has joined #aichallenge 2011-11-10T21:23:18 *** JorgeB has quit IRC (Ping timeout: 240 seconds) 2011-11-10T21:23:24 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-10T21:24:27 *** Migi32 has quit IRC (Remote host closed the connection) 2011-11-10T21:26:20 *** RangerMauve is now known as ChrisHanson 2011-11-10T21:26:33 *** ChrisHanson is now known as RangerMauve 2011-11-10T21:26:57 *** Antimony has joined #aichallenge 2011-11-10T21:28:19 @rainbow RangerMauve 2011-11-10T21:28:19 thestinger: RangerMauve 2011-11-10T21:28:29 *** alc has joined #aichallenge 2011-11-10T21:29:06 Y u trolling me? 2011-11-10T21:29:17 Bonus points for effort, though. 2011-11-10T21:29:49 *** dorisabayon has quit IRC (Ping timeout: 240 seconds) 2011-11-10T21:29:54 *** nplus has quit IRC (Read error: Connection reset by peer) 2011-11-10T21:31:10 *** amstan has quit IRC (Read error: Operation timed out) 2011-11-10T21:31:13 *** dorisabayon has joined #aichallenge 2011-11-10T21:32:56 *** amstan has joined #aichallenge 2011-11-10T21:32:56 *** ChanServ sets mode: +o amstan 2011-11-10T21:33:05 In the api, if there is an ant on a hill do we get both lines listed? 2011-11-10T21:35:08 *** bmh has joined #aichallenge 2011-11-10T21:35:09 *** bmh has joined #aichallenge 2011-11-10T21:35:56 *** amstan has quit IRC (Remote host closed the connection) 2011-11-10T21:36:25 *** amstan has joined #aichallenge 2011-11-10T21:36:25 *** ChanServ sets mode: +o amstan 2011-11-10T21:36:30 *** dom7b5 has quit IRC (Ping timeout: 240 seconds) 2011-11-10T21:36:30 hey amstan 2011-11-10T21:36:36 hey 2011-11-10T21:37:12 *** nplus has joined #aichallenge 2011-11-10T21:38:08 <_flag> janzert, amstan: Should I care that the C starter pull request to add hills doesn't record the owner of a dead ant? I'm not sure anyone really uses that information anyway, but it might be a little strange if someone did decide to use it and it wasn't there 2011-11-10T21:38:33 _flag: feel free to request such things 2011-11-10T21:38:55 *** dorisabayon has quit IRC (Ping timeout: 258 seconds) 2011-11-10T21:40:19 *** dorisabayon has joined #aichallenge 2011-11-10T21:40:54 *** twymer has quit IRC (Ping timeout: 276 seconds) 2011-11-10T21:42:21 *** twymer has joined #aichallenge 2011-11-10T21:43:51 Do we get dead ant info if it dies outside of your view radius? 2011-11-10T21:44:17 *** treeform has joined #aichallenge 2011-11-10T21:44:22 err, I mean if my own ant dies all by itself and I can no longer see that area, do I still see the dead ant line? 2011-11-10T21:45:04 *** rajanaresh has joined #aichallenge 2011-11-10T21:46:34 _flag: or if you feel like accept the pull request then add it yourself 2011-11-10T21:47:25 *** twymer has quit IRC (Ping timeout: 240 seconds) 2011-11-10T21:49:27 *** Jak_o_Shadows has joined #aichallenge 2011-11-10T21:50:01 *** dorisabayon has quit IRC (Ping timeout: 255 seconds) 2011-11-10T21:53:03 what do I have to do to make 'ant paste' recognize that I'm pasting a map and not a reply? 2011-11-10T21:53:52 *** Antimony has quit IRC (Ping timeout: 258 seconds) 2011-11-10T21:56:29 *** dorisabayon has joined #aichallenge 2011-11-10T21:59:59 *** dvladim has joined #aichallenge 2011-11-10T22:01:04 *** dorisabayon has quit IRC (Ping timeout: 248 seconds) 2011-11-10T22:01:26 *** TheLinker has quit IRC (Ping timeout: 252 seconds) 2011-11-10T22:01:59 *** dorisabayon has joined #aichallenge 2011-11-10T22:07:28 *** dvladim has quit IRC (Ping timeout: 245 seconds) 2011-11-10T22:11:56 *** dorisabayon has quit IRC (Ping timeout: 260 seconds) 2011-11-10T22:12:34 *** Antimony has joined #aichallenge 2011-11-10T22:12:43 *** dorisabayon has joined #aichallenge 2011-11-10T22:15:20 *** Rinum has quit IRC (Ping timeout: 265 seconds) 2011-11-10T22:19:33 *** dorisabayon has quit IRC (Ping timeout: 258 seconds) 2011-11-10T22:19:36 *** hop41 has joined #aichallenge 2011-11-10T22:20:38 *** dorisabayon has joined #aichallenge 2011-11-10T22:22:27 *** Jak_o_Shadows has quit IRC (Remote host closed the connection) 2011-11-10T22:26:19 *** BenJackson_ has joined #aichallenge 2011-11-10T22:27:05 *** dorisabayon has quit IRC (Ping timeout: 260 seconds) 2011-11-10T22:27:33 Extrarius: reply looks like json, maps look like plain text(start with rows or col i think) 2011-11-10T22:27:56 *** dorisabayon has joined #aichallenge 2011-11-10T22:30:29 *** Jak_o_Shadows has joined #aichallenge 2011-11-10T22:31:08 *** rajanaresh has quit IRC (Ping timeout: 258 seconds) 2011-11-10T22:32:09 *** TheLinker has joined #aichallenge 2011-11-10T22:36:29 *** twymer has joined #aichallenge 2011-11-10T22:38:59 *** chris__0076 is now known as chris_0076 2011-11-10T22:44:54 *** kiv has quit IRC (Ping timeout: 240 seconds) 2011-11-10T22:45:39 *** hjax has joined #aichallenge 2011-11-10T22:47:00 *** hjax has joined #aichallenge 2011-11-10T22:48:02 *** xbelt has joined #aichallenge 2011-11-10T22:52:15 *** Jak_o_Shadows has quit IRC (Remote host closed the connection) 2011-11-10T22:52:32 *** xbelt has quit IRC (Ping timeout: 260 seconds) 2011-11-10T22:53:26 *** kiv has joined #aichallenge 2011-11-10T22:56:46 *** dvladim has joined #aichallenge 2011-11-10T22:59:26 good way to win http://aichallenge.org/visualizer.php?game=83616&user=164 2011-11-10T23:01:52 *** djr_ has joined #aichallenge 2011-11-10T23:03:46 *** Cyndre has quit IRC (Quit: Leaving) 2011-11-10T23:05:02 *** djr_ has quit IRC (Read error: Connection reset by peer) 2011-11-10T23:05:49 *** djr_ has joined #aichallenge 2011-11-10T23:13:55 *** hop41 has quit IRC (Quit: Page closed) 2011-11-10T23:17:45 *** dorisabayon has quit IRC (Ping timeout: 276 seconds) 2011-11-10T23:18:25 *** dorisabayon has joined #aichallenge 2011-11-10T23:21:27 *** Jak_o_Shadows has joined #aichallenge 2011-11-10T23:23:29 *** sai has joined #aichallenge 2011-11-10T23:24:04 can I cancel my account? 2011-11-10T23:24:40 how can i upload stuff in c++11 ? 2011-11-10T23:24:52 name it .cpp, I think 2011-11-10T23:25:19 *** dvladim has quit IRC (Ping timeout: 248 seconds) 2011-11-10T23:25:56 yeah, change MyBot.cc to MyBot.cpp and -std=c++0x will be used 2011-11-10T23:26:06 ok 2011-11-10T23:26:15 and keep in mind that the server only has gcc 4.5 2011-11-10T23:26:18 http://gcc.gnu.org/projects/cxx0x.html 2011-11-10T23:26:19 not many people doing it judging from the language page 2011-11-10T23:27:10 *** Rinum has joined #aichallenge 2011-11-10T23:27:42 how are you guys handling exploration? 2011-11-10T23:28:09 Rinum: Make unexplored squares attractive targets? 2011-11-10T23:28:42 nickjohnson: but it's too expensive to loop though the entire map to make those attractive? 2011-11-10T23:28:49 I beat up lone ants from other players and make them tell me what they know 2011-11-10T23:28:54 Rinum: Well, that depends on your overall strategy 2011-11-10T23:28:56 with a little ant dance 2011-11-10T23:29:10 Using Collaborative Diffusion or a similar strategy that relies on marking up map squares with values, you're already doing that. 2011-11-10T23:29:13 nickjohnson: let's say with CD 2011-11-10T23:29:17 hm, now I'm pondering how many bits of information two ants could reasonably communicate during a game 2011-11-10T23:29:26 already doing that? how so? 2011-11-10T23:29:35 Rinum: Right, so you make the value of unexplored squares something higher than explored squares and diffuse them 2011-11-10T23:29:43 You're already looping over the whole area to diffuse 'scents' 2011-11-10T23:29:46 *** JorgeB has joined #aichallenge 2011-11-10T23:30:22 nickjohnson: by 'unexplored' do you mean squares ants haven't 'touched' (not just visible) 2011-11-10T23:30:23 In mine, I made squares outside line-of-sight emit a 'food' smell at about half the level of actual food 2011-11-10T23:30:34 Rinum: Which you use is up to you 2011-11-10T23:30:43 I do that too, but my bot times out 2011-11-10T23:30:46 Depends if you want good live coverage of the map, or just want to explore the whole thing at least once 2011-11-10T23:30:52 What language are you using? 2011-11-10T23:30:55 C++ 2011-11-10T23:31:04 You must be doing something fairly slow, then 2011-11-10T23:31:12 I can do a few iterations per turn over the whole map in Python 2011-11-10T23:31:18 Though I'm thinking of moving to Go so I can do more 2011-11-10T23:31:29 Have you tried profiling your code? 2011-11-10T23:31:37 well I'm trying to get the scent to travel the entire map... that could be why 2011-11-10T23:32:16 my C++ bot only took about 5ms to run when it was doing only something CD like 2011-11-10T23:32:26 Using BFS, or just by iterating multiple times? 2011-11-10T23:32:37 *** Jak_o_Shadows has quit IRC (Remote host closed the connection) 2011-11-10T23:32:41 With typical CD, you just do some number of diffusion steps per game step, and keep the diffusion map between turns 2011-11-10T23:32:55 nickjohnson: BFS 2011-11-10T23:32:55 I re-evaluate everything every turn 2011-11-10T23:33:04 but it's not really CD 2011-11-10T23:33:16 i've been re-evaluating each turn 2011-11-10T23:33:26 Even then you should be able to do something fairly quickly. Your BFS can prune branches that encounter another unexplored square, or one with a shorter distance to an unexplored squarew. 2011-11-10T23:33:27 a correct BFS should be equivalent in cost to just iterating once across the map 2011-11-10T23:33:38 BenJackson_: Yes, but per source, not in total. 2011-11-10T23:33:57 you can do all equivalent sources in one traversal 2011-11-10T23:34:16 Hm, I suppose you can. That's not really BFS, but the effect is the same. 2011-11-10T23:34:24 sure it is 2011-11-10T23:34:29 the root is just not a grid square 2011-11-10T23:34:34 it's some imaginary concept like "food" 2011-11-10T23:34:44 Fair enough 2011-11-10T23:34:46 all of the children of that root are your first queue 2011-11-10T23:34:51 Hylaean Breadth First Search ;) 2011-11-10T23:35:17 AKA Platonic - (http://anathem.wikia.com/wiki/Hylaean_Theoric_World) 2011-11-10T23:36:08 I like to keep my bot fast because it's quicker to iterate 2011-11-10T23:36:16 my first python version was annoyingly slow to debug 2011-11-10T23:36:30 and timed out without careful time controls 2011-11-10T23:36:35 This is why I'm thinking of switching to Go 2011-11-10T23:36:44 As much as I love Python, it's not up to scratch speed-wise for this challenge 2011-11-10T23:36:49 the C++ version has probably never taken more than 25ms to compute a turn even with 250 ants 2011-11-10T23:37:23 actually the early versions were probably O(map size) and didn't care about ant counts at all 2011-11-10T23:37:25 * nickjohnson is also pretty sure the Python starter kit was written by someone who isn't that familiar with Python, which is odd since the server is written in Python 2011-11-10T23:37:39 BenJackson_: Well, you still have to iterate over the ants to give them orders 2011-11-10T23:37:42 the server is not in the most fluent python either 2011-11-10T23:37:48 *** Jak_o_Shadows has joined #aichallenge 2011-11-10T23:38:29 sure, but board size is like 150*150 or ~ 22000 and you have <100 ants most of the time 2011-11-10T23:38:30 is there a way to change the turn time? 2011-11-10T23:38:53 playgame.py has a ton of options 2011-11-10T23:39:01 including --turntime and --loadtime 2011-11-10T23:39:38 nickjohnson: I meant more in constrast now when my bot uses algorithms that are n^2 in ants or worse 2011-11-10T23:39:52 ouch 2011-11-10T23:40:00 I benchmarked a n^3 algorithm that would have been fine for 300+ ants 2011-11-10T23:40:38 Still doing CD stuff, or something entirely different? 2011-11-10T23:41:04 no, that was an entirely different idea which I didn't end up using 2011-11-10T23:42:06 *** Jak_o_Shadows has quit IRC (Remote host closed the connection) 2011-11-10T23:42:09 I'm not really doing CD at all, just something similar 2011-11-10T23:42:13 I hadn't heard of CD when I started 2011-11-10T23:43:30 seems like everyone is doing CD now x.x 2011-11-10T23:44:42 hopefully there won't be a similar revelation in combat 2011-11-10T23:44:52 right now combat is what stratifies the top tier 2011-11-10T23:46:01 *** Jak_o_Shadows has joined #aichallenge 2011-11-10T23:46:11 yup 2011-11-10T23:47:59 also if it turns out combat is as easy as CD-like structures are for overall strategy I'm going to feel really dumb 2011-11-10T23:50:16 lol ditto 2011-11-10T23:50:50 i bet xathis' strategy is so simple 2011-11-10T23:51:05 *** treeform has quit IRC (Ping timeout: 260 seconds) 2011-11-10T23:52:18 *** yoden has quit IRC (Quit: Leaving.) 2011-11-10T23:52:28 is CD really that popular? it seemed like something that wouldn't hold up well at high tiers 2011-11-10T23:52:54 there's still a ton of refinements to what you put INTO your "CD-like" algorithm 2011-11-10T23:53:09 starter packages don't even remember hills, for example 2011-11-10T23:53:51 after my reloaded bot played some ~500 level bots I got curious about what a "low sigma" 500-600th place bot was like 2011-11-10T23:53:56 so I watched some games 2011-11-10T23:53:58 aichallenge: = epsilon * r0d1a37f / (3 files): add octave starter package - http://git.io/iOf8Sw 2011-11-10T23:53:59 aichallenge: = epsilon * r15813ac / worker/compiler.py : add octave source detection - http://git.io/m8msXA 2011-11-10T23:54:00 aichallenge: = epsilon * r62aadd1 / setup/worker_setup.py : include octave3.2 ubuntu package in the list of distribution languages to be installed - http://git.io/cznPvA 2011-11-10T23:54:01 aichallenge: = epsilon * r0abcef2 / website/starter_packages.php : update starter packages webpage to include octave package - http://git.io/TYkRmA 2011-11-10T23:54:02 aichallenge: Janzert epsilon * r493cc98 / (6 files in 4 dirs): 2011-11-10T23:54:02 aichallenge: Merge pull request #364 from pedrosorio/epsilon 2011-11-10T23:54:03 aichallenge: Octave starter package - http://git.io/468AbA 2011-11-10T23:54:25 *** TheLinker has quit IRC (Remote host closed the connection) 2011-11-10T23:54:44 one funny moment involved a bot which found a hill, sent a huge force to cap it, and didn't give up until it stepped on the square where the hill was, even though it was captured shortly after first being sighted 2011-11-10T23:54:51 meanwhile that same bot got ninja capped at his base 2011-11-10T23:55:40 it's really the combat strat that creates the rankings 2011-11-10T23:55:54 *** RangerMauve has quit IRC (Read error: Connection reset by peer) 2011-11-10T23:56:02 I kind of want to do a meta-analysis of the whole contest to see where the breaks are 2011-11-10T23:56:12 eg what rank a food-seeking bot coould get 2011-11-10T23:56:14 lol I've been doing that 2011-11-10T23:56:24 one that seeks food and suicides at hills 2011-11-10T23:56:29 one with combat, etc 2011-11-10T23:56:39 combat seems to be top-25 right now 2011-11-10T23:56:43 *** Antimony has quit IRC (Ping timeout: 256 seconds) 2011-11-10T23:56:48 skill ~30s solely food gathering 2011-11-10T23:56:52 certainly I looked at some 50-60 level bots that appeared to be oblivious 2011-11-10T23:56:57 ~40s very basic combat 2011-11-10T23:57:17 ~50s slightly more basic combat (much like what you described) 2011-11-10T23:57:22 etc. 2011-11-10T23:57:26 Rinum: I was thinking of actually writing a program to consume replays and do an analysis 2011-11-10T23:57:39 sounds tough 2011-11-10T23:57:53 *** Jak_o_Shadows has quit IRC (Remote host closed the connection) 2011-11-10T23:57:57 you could easily measure how many combat events came out numerically in favor of a bot 2011-11-10T23:58:11 how rapidly it feeds, etc 2011-11-10T23:58:39 maybe if I were currently dominating the competition I would think about doing something like that x.x but for now... must improve the bot! 2011-11-10T23:59:39 I'm kind of surprised my non-CD, slow python bot made it to the 60s 2011-11-10T23:59:42 http://aichallenge.org/profile.php?user=3469 2011-11-10T23:59:50 your python bot has a higher skill than mine did before I abandoned it :)