2011-11-11T00:00:00 lol 2011-11-11T00:00:16 mine hit 57.64 2011-11-11T00:00:19 have you made any submissions after that? 2011-11-11T00:00:21 lots of ants got no orders due to timeouts and stuff 2011-11-11T00:00:22 I got to rank 27 with my previous upload - it had no combat or awareness of enemy ants 2011-11-11T00:00:33 it just took ~a week to climb up from rank 50 2011-11-11T00:00:39 my previous one hit 78.19 2011-11-11T00:00:44 nice! 2011-11-11T00:00:45 *** Jak_o_Shadows has joined #aichallenge 2011-11-11T00:00:47 which was variously 20-25 2011-11-11T00:01:02 my new one slaughters my old one in local game tests 2011-11-11T00:01:17 but I only uploaded last night so I'm not getting competitive games yet 2011-11-11T00:01:43 I did lose one but I can't learn much from it: rankings were almost totally luck in that game 2011-11-11T00:02:06 your bot looks great! 2011-11-11T00:02:56 it has been really interesting to work on it 2011-11-11T00:03:15 my early ideas about combat were so wrong 2011-11-11T00:03:19 sooo wrong 2011-11-11T00:03:28 lol mine too 2011-11-11T00:04:12 I'd love to run a bot with a lot of tweakable knobs, like CD, through a simulated annealing process to find the best settings 2011-11-11T00:04:13 I'm definitely glad I ditched python though 2011-11-11T00:04:17 I'm betting it would make a big difference 2011-11-11T00:04:53 Rinum: I ditched python also and it's so much easier now 2011-11-11T00:05:04 I did too :) 2011-11-11T00:05:20 using C++ for this makes it much easier imo, because you barely have to worry about performance 2011-11-11T00:05:30 yup 2011-11-11T00:05:43 with python I would have had to make a bunch of compromises or figure out much better algorithms than what I'm using 2011-11-11T00:05:48 unless you're using an n^4 algorithm x.x 2011-11-11T00:06:28 not worrying about timeouts is nice, but simulating 750 turn games in 20 seconds is fantastic 2011-11-11T00:06:38 I don't have real coordinated combat yet, so the time is basically independent of the number of ants 2011-11-11T00:08:15 i also am revamping my combat. i have some promising simulations now but i have no idea if i can scale this 2011-11-11T00:08:35 early I didn't worry abuot suiciding into other bots 2011-11-11T00:08:49 then I realized that in an N-way game if I cancel out one other guy we both just went -1 vs everyone else 2011-11-11T00:08:51 yeah, you need to make some sort of tradeoff 2011-11-11T00:08:58 so it turns out just "not suiciding" is the first big combat step 2011-11-11T00:09:02 heh. 2011-11-11T00:09:10 a 2 for 1 ratio might be good. 2011-11-11T00:09:25 yeah, now that I have combat I don't worry about it as much 2011-11-11T00:09:43 but early on I dismissed it as "oh I'll just feed fast and crash into things" 2011-11-11T00:09:55 then people developed the "square defense" 2011-11-11T00:10:04 it's pretty funny to watch blind bots crash into that 2011-11-11T00:10:04 but yeah my bot is combat aware to a point, but is unable to coordinate at all, so will still send hundreds of ants in a single-file column into certain death 2011-11-11T00:10:08 yep 2011-11-11T00:10:20 square defense? 2011-11-11T00:10:30 an ant on each corner of a hill 2011-11-11T00:10:32 http://aichallenge.org/visualizer.php?game=81875&user=3469 2011-11-11T00:10:32 my ants basically just strafe/encircle the enemy until they aren't outnumbered, there's no coordination 2011-11-11T00:10:42 that's one of Rinum's games, look at "wilsone" 2011-11-11T00:10:56 haha, cool idea 2011-11-11T00:10:58 I allow 1 for 1 trades though, so it doesn't work very well atm :P 2011-11-11T00:11:07 once I figured out how to do "ant on each corner" I realized I knew how to do that "GG" thing one guy was doing 2011-11-11T00:11:18 heh, yeah. that was awesome. 2011-11-11T00:11:29 that's temping to add. 2011-11-11T00:11:50 "ant on each corner"? 2011-11-11T00:11:57 a a 2011-11-11T00:11:58 o 2011-11-11T00:11:58 I'm not sure if "square defense" took off on aichallenge.org but it was an early fluxid innovation and pretty soon everyone had it 2011-11-11T00:11:59 a a 2011-11-11T00:12:00 yeah, the static X defense formation 2011-11-11T00:12:04 a being ant, o being hill 2011-11-11T00:12:05 Of your hive? 2011-11-11T00:12:07 ah 2011-11-11T00:12:18 So what's the "GG thing", then? 2011-11-11T00:12:20 I see people using a + formation too, but it doesn't work as well 2011-11-11T00:12:25 I can save you some time if you want to watch my v5 on a multi-hill maze to see how NOT to do it 2011-11-11T00:12:32 I was letting it burn way too many ants on defense 2011-11-11T00:12:54 nickjohnson: some bot spells out "GG" In ants when it has a large surplus 2011-11-11T00:12:59 hah, nice 2011-11-11T00:13:15 Hm, the main problem with defensive formations would be making sure you don't form them until you have enough ants 2011-11-11T00:13:23 Otherwise I'd just make them very attractive hill-climbing spots :) 2011-11-11T00:13:23 as I said 2011-11-11T00:13:35 BenJackson_: lmao, have an example of the GG in action? 2011-11-11T00:13:41 yeah, I've left out defense so far - I'm trying out various ideas though 2011-11-11T00:13:48 Rinum: I don't remember who it was 2011-11-11T00:13:49 adding the square formation thing ruined my bot :P 2011-11-11T00:15:28 argh it's hard to find multi-hill mazes by name in the game list 2011-11-11T00:17:09 *** u_ has joined #aichallenge 2011-11-11T00:18:06 xathis is really holding onto his lead. i was sure he'd be unseated by now 2011-11-11T00:18:34 yeah, it's amazing 2011-11-11T00:18:59 he's still FAR ahead of everyone 2011-11-11T00:19:01 did his skill just break 90? 2011-11-11T00:19:33 it's broken 90 before... happened again recently 2011-11-11T00:19:40 <_flag> Although if you look at the mus there are a few people that are somewhat close 2011-11-11T00:19:48 <_flag> Just with higher sigmas 2011-11-11T00:19:48 mus? 2011-11-11T00:19:52 <_flag> GreenTea, for example 2011-11-11T00:19:58 <_flag> However over the skill 2011-11-11T00:20:18 <_flag> The mu is the estimated rating, and the sigma is the certaintity of that rating 2011-11-11T00:20:25 <_flag> The skill is the mu - sigma*3 2011-11-11T00:20:48 i worry that his sigma is artificially low because of the long history 2011-11-11T00:21:04 <_flag> Possibily 2011-11-11T00:21:33 <_flag> Losing games like this is extremely frustrating: http://aichallenge.org/visualizer.php?game=83510&user=31 2011-11-11T00:21:35 what's scarier is that his bot has obvious bugs 2011-11-11T00:21:42 how can you see mu and sigma? 2011-11-11T00:21:45 so he could definitely swoop in with big improvements if needed 2011-11-11T00:21:48 u_: hover over "score" 2011-11-11T00:21:50 <_flag> BenJackson_: like what? 2011-11-11T00:21:53 skill 2011-11-11T00:22:10 BenJackson_: he's definitely improving his bot 2011-11-11T00:22:20 _flag: well I just watched one of his recent games where he nearly stalemated despite having 500+ ants because he didn't push his advantage 2011-11-11T00:22:32 one of his hills was bottled up by a bot which timed out leaving a "plug" that he didn't go through 2011-11-11T00:22:48 i've never seen a bot break through a plug like that 2011-11-11T00:22:50 his combat clearly is of the type which gives a fixed output for a fixed input 2011-11-11T00:22:57 u_ oh, mine would 2011-11-11T00:23:04 my combat is sort of sloppy-aggressive 2011-11-11T00:23:20 <_flag> Ah, yeah I've seen that too. That's true of most of the top because because they're left over from the beta where engaging in combat of any kind where victory was not certain was bad because there were no hills to worry about :) 2011-11-11T00:23:40 in fact if you watch one of my recent games one of the bots times out and I do eventually cap it despite the huge clog of ants it leaves 2011-11-11T00:24:06 _flag: AHHH I had wondered why the top ants were so super conservative about fighting 2011-11-11T00:24:33 my algorithm just doesn't capture the idea of defense very well :) 2011-11-11T00:24:55 *** Antimony has joined #aichallenge 2011-11-11T00:25:01 <_flag> Even still though, being conservative about fighting is still the best approach in the majority of situations 2011-11-11T00:25:17 somewhat, yes 2011-11-11T00:25:19 <_flag> Simply because the game is a free-for-all multiplayer type 2011-11-11T00:25:34 <_flag> And it's beneficial to allow your opponents to fight each other while you gain a massive advantage 2011-11-11T00:25:58 on the other hand you can't make an omlatte without breaking eggs :) 2011-11-11T00:26:03 *e 2011-11-11T00:26:16 *** Areks has joined #aichallenge 2011-11-11T00:26:21 I was disappointed to see that number of players is a secret 2011-11-11T00:26:25 <_flag> Obviously it's a balence, but I think the results show that xathis and the other top bots are doing it correctly :) 2011-11-11T00:26:33 makes it hard to figure out what kind of map experience you're getting into 2011-11-11T00:26:43 *** grom358 has joined #aichallenge 2011-11-11T00:26:53 strcat how many versions of the bot do u have? 2011-11-11T00:27:02 grom358: not many 2011-11-11T00:27:08 OH you're strcat 2011-11-11T00:27:13 every time I add a feature I make a branch in git, and test it on the tcp server 2011-11-11T00:27:27 some things like defense ended up making the bot worse 2011-11-11T00:27:31 I found something last night I really want to try now 2011-11-11T00:27:34 sometimes I look at fluxid and my bot is playing every game against strcat version somethingorother 2011-11-11T00:27:38 so they're still branches :) 2011-11-11T00:27:45 Cooperative A* pathfinding :) 2011-11-11T00:27:53 <_flag> strcat is the bot that most frequently kills me on tcp simply by rushing my hill with all its ants :P 2011-11-11T00:28:04 my fluxid bot is old 2011-11-11T00:28:20 thestinger: your bots are good! 2011-11-11T00:28:21 I debated updating it but I figure instant death is not that useful to people trying to get started 2011-11-11T00:28:37 I still have zero combat lol 2011-11-11T00:28:38 their stupidity gets them points :) 2011-11-11T00:28:42 I keep playing around with pathfinding 2011-11-11T00:29:09 improving my movement made up for my lack of combat a lot 2011-11-11T00:29:27 yeah.. i need some sort of flocking or something on the enemy hill 2011-11-11T00:29:36 that would sort out a lot for me 2011-11-11T00:29:54 I already have avoid code when gathering 2011-11-11T00:30:05 just need something little smarter for attacking a hill 2011-11-11T00:30:17 instead of going in line towards it 2011-11-11T00:30:19 thestinger: nice thing about a bunch of git branches is also having a stable of local bots to try new versions against 2011-11-11T00:31:08 yeah I still haven't setup git yet lol.. will be soon though 2011-11-11T00:31:24 BenJackson_: yeah, having a version of my bot that does defense is very useful for testing 2011-11-11T00:32:01 the one pre-combat version I keep around is getting too weak at this point 2011-11-11T00:32:19 thestinger: without giving away too much of your secrets.. how exactly are u doing combat? I haven't thought of a way to deal with it yet 2011-11-11T00:32:19 it is simulating the imbalance that results from having a weak neighbor vs a strong neighbor 2011-11-11T00:33:12 I haven't got the foggest idea yet on how to get multiple ants to coordinate an attack 2011-11-11T00:33:33 grom358: I'm not really doing coordinated combat 2011-11-11T00:33:37 it's just movement code 2011-11-11T00:34:10 I'm going to switch to doing something like putting all orders on a queue and altering them for combat 2011-11-11T00:34:17 is the movement coordinated? so you have a wall on the front line? 2011-11-11T00:34:54 sort of, they'll try to form a wall when they reach enemies 2011-11-11T00:35:01 until then it's pathfinding/movement 2011-11-11T00:35:30 so do you have a structure/class for forming walls? so the wall can move as one? 2011-11-11T00:35:34 thestinger: care to point out a fluxid (or aichallenge) replay that shows the state of the art in strcat_* bots? 2011-11-11T00:36:37 grom358: my tip would be that forming walls and stuff is better produced as emergent behavior of a general combat algorithm than by trying to do it explicitly 2011-11-11T00:38:06 BenJackson: yeah.. that is what I was hoping todo. Somehow get a number of ants to be focusing on a cell at the same time 2011-11-11T00:38:38 BenJackson_: hmm, hard to find a good game - strcat_next and strcat_gatherzone are basically my current codebase though 2011-11-11T00:38:55 thestinger: oh you're strcat? 2011-11-11T00:39:04 yeah :) 2011-11-11T00:39:09 grom358: there is (IMO) an element of "aha" puzzle to ant combat 2011-11-11T00:39:13 strcat_combat is like 10 times stronger than strcat_gather 2011-11-11T00:39:29 I really would like to discuss it but I fear it is seriously against my interests 2011-11-11T00:39:40 strcat_forage is just to amuse me - it basically just forages and disregards everything else 2011-11-11T00:39:48 BenJackson_: haha, same here 2011-11-11T00:39:52 BenJackson: fair enough 2011-11-11T00:40:05 thestinger: That'd be great if winners were determined by numbers ;) 2011-11-11T00:40:27 I've wondered how well a bot that attempts to distribute itself evenly over the map would do. Perhaps in groups of 4 to make it difficult to wipe out the 'nodes'. 2011-11-11T00:40:33 Probably not wonderfully, but it'd be amusing at least. 2011-11-11T00:41:33 every damn time I play new games on fluxid I'm reminded of how food-poor it is 2011-11-11T00:41:39 and somehow every time I forget about it 2011-11-11T00:41:42 speak of the devil.. I just played strcat bot again :) 2011-11-11T00:41:51 http://ants.fluxid.pl/replay.29800 this basically shows you what my bots do when they don't get wiped out early on due to lack of defense 2011-11-11T00:42:10 it works well on mazes, but they suck at the random walk maps 2011-11-11T00:42:18 my defense branch does a lot better on those 2011-11-11T00:42:48 lol.. I have token defence.. I post ants on the corners of my hills 2011-11-11T00:43:06 seems to stop a lot of attacks 2011-11-11T00:43:19 thestinger: I love watching games on that map to see how ants handle that cul-de-sac 2011-11-11T00:43:23 except people who have got combat code going 2011-11-11T00:43:35 *** ybit has quit IRC (Quit: leaving) 2011-11-11T00:43:50 *** hacklash has joined #aichallenge 2011-11-11T00:44:26 my bot is great at pathfinding especially when they don't get stuck on each other 2011-11-11T00:44:46 Cooperative A* will fix the stuckness 2011-11-11T00:44:51 that sometimes happens 2011-11-11T00:46:22 mmm.. except I using BFS for a lot of the pathfinding. mmmm 2011-11-11T00:48:02 actually BFS makes up probably like 95% of the pathfinding 2011-11-11T00:49:24 *** bmh has quit IRC (Quit: bmh) 2011-11-11T00:50:51 :'( get raped by bots with combat code 2011-11-11T00:51:21 *** rofer is now known as Rofer 2011-11-11T00:51:28 my development basically ground to a halt for several days while I pondered combat 2011-11-11T00:52:08 yeah.. looking at my current bot that I built today... I think I need to work on combat over improvements in pathfinding 2011-11-11T00:52:37 like getting ants to fight other ants with minimal loses 2011-11-11T00:53:10 *** praveen_ has joined #aichallenge 2011-11-11T00:53:19 my bot runs faster when there are more ants in the field x.x 2011-11-11T00:53:34 my bot speeds up a lot as it explores 2011-11-11T00:53:42 maybe the same thing is going on with yours? 2011-11-11T00:53:46 I have an order queue... but I need someway to have orders that are only issued with other orders.. like batch orders that are all or nothing 2011-11-11T00:54:18 yup 2011-11-11T00:54:53 *** Rofer is now known as rofer 2011-11-11T00:54:58 is it possible your ant actually runs faster when it has seen more of the map? 2011-11-11T00:55:12 the default fill type is "LAND" (at least in the python and c++ starters) 2011-11-11T00:55:27 so the less you've seen the more willing your pathfinding might be to flood fill the entire map area 2011-11-11T00:55:56 I'm using CD on all the nonvisible parts 2011-11-11T00:56:07 how exactly does that work? My bot assumes a tile is land if it hasn't been seen yet.. So as I explore I don't have to reculate paths as much.. but doesn't really get faster AFAIK 2011-11-11T00:56:21 which is very expensive with my code 2011-11-11T00:56:35 my python bot saved paths 2011-11-11T00:56:35 oh.. I see.. yeah less to search on maze maps as it finds water tiles 2011-11-11T00:56:37 my c++ bot doesn't 2011-11-11T00:56:39 I don't bother calculating paths anymore 2011-11-11T00:56:47 I don't use paths anymore 2011-11-11T00:57:12 I think combat is the main difficulty in writing a good bot 2011-11-11T00:57:19 how do you move then? just random walk? 2011-11-11T00:57:21 I'm trying to figure out how to implement some combat in mine 2011-11-11T00:57:43 I build up a distance map and just use gradient descent 2011-11-11T00:57:47 well 2011-11-11T00:57:51 sort of 2011-11-11T00:58:00 yep, that's what I did when I switched to c++ 2011-11-11T00:58:03 ebasically I'm using A* 2011-11-11T00:58:08 out of paranoia about how slow the python had been 2011-11-11T00:58:16 lol 2011-11-11T00:58:18 but I recalcualte everything each tick 2011-11-11T00:58:26 (first version took like 5ms/turn) 2011-11-11T00:59:25 Darn. my current bot seems to be doing badly 2011-11-11T00:59:35 it's still at skill 50 after 5 games 2011-11-11T01:00:00 heh, I was trying to project where my current bot will go by comparing its skill trajectory 2011-11-11T01:00:08 TIL ants can kill ants across water x.x 2011-11-11T01:00:20 the problem with A*.. say you see food. How do you determine the closest ant to said food? A* needs start and end points 2011-11-11T01:00:30 then I realized you can just hover "score" and see mu directly 2011-11-11T01:00:56 A* does, but A* is just a refinement of BFS to avoid searching as broadly 2011-11-11T01:01:09 I can tell you for sure that it's not necessary if you're working in C++ 2011-11-11T01:01:23 and (though I never really tried to fix it) ieven A* is not fast enough for python 2011-11-11T01:01:31 *** sai has quit IRC (Quit: Page closed) 2011-11-11T01:01:52 I use BFS for food fetching cause not sure which ant will be closer without searching in all directions 2011-11-11T01:02:43 *** Palmik has joined #aichallenge 2011-11-11T01:02:49 ACtually 2011-11-11T01:03:02 I don't bother finding the closest ant ot each food 2011-11-11T01:03:20 I do have anticlustering to reduce the frequency of multiple ants goign to teh same food though 2011-11-11T01:03:35 the only time I use A* is for recalculating paths from the BFS search.. like say the BFS path included an unexplored water tile.. then I drop the path and use A* to the target 2011-11-11T01:04:14 so yeah.. in the end A* is used very little in my bot 2011-11-11T01:04:24 personally I think anyone who is trying to remember a given ant from one turn to another is probably making a mistake 2011-11-11T01:04:41 Currently I remember ants but I don't use the info for anything 2011-11-11T01:05:01 BenJackson: I do it to save the paths 2011-11-11T01:05:07 Well I remember my ants. Enemies are a lot harder 2011-11-11T01:05:27 I used to save paths, but recalculating each turn is easier and more effective 2011-11-11T01:05:38 and sometimes I getting random spikes in performance of my Java bot... if in C++ could probably just recalculate each turn 2011-11-11T01:05:42 OJ is such a killstealer. http://aichallenge.org/visualizer.php?game=83854&user=3845 2011-11-11T01:05:52 *** pairofdice has joined #aichallenge 2011-11-11T01:06:40 I have a bot that I was working on that didn't save paths but couldn't get it working 2011-11-11T01:07:05 Switch to C++ 2011-11-11T01:07:22 And then you won't have to worry about Java's awfull syntax either 2011-11-11T01:07:34 Java syntax no worse then C++ really 2011-11-11T01:07:41 it's a lot more verbose 2011-11-11T01:07:56 depends 2011-11-11T01:08:10 they pretty similar in verbose to me 2011-11-11T01:08:13 I was going to use Scala but then I saw that post that said their JRE sandboxing for the contest had performance issues 2011-11-11T01:08:15 foo.bar += 3 vs foo.setBar( foo.getBar() + 3)? 2011-11-11T01:08:17 sooo I said fuckit and used C++ 2011-11-11T01:08:38 actually I've improved my knowledge of STL doing this which is nice 2011-11-11T01:08:45 take something like declaring a list of integers 2011-11-11T01:08:45 I learned C++ for this :) 2011-11-11T01:08:56 in C++, you just do std::Vector foo; 2011-11-11T01:09:09 in Java you have List foo = new ArrayList(); 2011-11-11T01:09:34 Antimony: doing two different things there.. your C++ example is declaring it on the stack 2011-11-11T01:09:41 the Java one is on the heap 2011-11-11T01:09:43 well JAva doesn't even give you the option 2011-11-11T01:09:52 most of the time you want it on the stack 2011-11-11T01:09:59 true 2011-11-11T01:10:04 Antimony: In that replay it almost looks like blu has a 50 ant threshold to change behavior 2011-11-11T01:10:17 but the JIT puts it on the stack a lot of the time 2011-11-11T01:10:44 grom358: actually the "control" information for std::vector is on the stack but its contents are allocated 2011-11-11T01:10:56 c++0x has std::array if you want it all on the stack 2011-11-11T01:11:08 yeah 2011-11-11T01:11:14 or boost::array if you've got an old compiler 2011-11-11T01:11:58 *** bmh has joined #aichallenge 2011-11-11T01:11:59 *** bmh has joined #aichallenge 2011-11-11T01:12:00 and it's even worse if you consider common coding practices as well as the syntax 2011-11-11T01:12:07 especially all that getter/setter cruft 2011-11-11T01:12:30 heh, my current aichallenge version on fluxid vs no less than THREE strcats: http://ants.fluxid.pl/replay.29819 2011-11-11T01:13:15 I've always had problems with TCP 2011-11-11T01:13:16 and they all failed :P 2011-11-11T01:13:24 it often randomly crashes after a while 2011-11-11T01:13:30 it leaks sockets 2011-11-11T01:13:30 with WindowsError permission denied 2011-11-11T01:13:52 is there anywhere you can get a fixed version? 2011-11-11T01:14:03 I don't think so 2011-11-11T01:14:13 the server is much worse than the client, fluxid has to keep restarting it 2011-11-11T01:14:26 it leaks a whole thread after each game or something :\ 2011-11-11T01:14:27 *** bmh has quit IRC (Client Quit) 2011-11-11T01:14:29 the C++ starter bot crashes if run with no input 2011-11-11T01:14:36 which the tcpclient will do if the server times out 2011-11-11T01:14:50 that's funny because the client ALSO leaks a thread on every game 2011-11-11T01:15:48 wow 2011-11-11T01:15:54 the version I have imports threading but doesn't use it 2011-11-11T01:16:06 it works so I haven't updated it 2011-11-11T01:16:21 watching my tcp client scroll I can tell if I'm playing against python bots 2011-11-11T01:16:29 how so? 2011-11-11T01:16:33 they're slow? 2011-11-11T01:16:36 yessss 2011-11-11T01:17:00 I should check with fluxid to see if he has the current aichallenge.org food settings 2011-11-11T01:17:09 food always seems scarce on fluxid games and I hate to optimzie for it 2011-11-11T01:17:33 Do you have special timeout code for tcp? 2011-11-11T01:17:40 That's another big difference from the server 2011-11-11T01:18:20 *** Jak_o_Shadows1 has joined #aichallenge 2011-11-11T01:18:21 I decided to make my bot pretend like the timeout is 500ms when playing on tcp 2011-11-11T01:19:23 shouldn't need to, the server gives you spare time for a similar reason 2011-11-11T01:20:19 *** Rinum has quit IRC (Quit: Page closed) 2011-11-11T01:20:55 *** Jak_o_Shadows has quit IRC (Ping timeout: 260 seconds) 2011-11-11T01:22:47 what do you mean? 2011-11-11T01:23:37 Anyway, my current goal is to implement combat, assuming all enemy ants are stationary 2011-11-11T01:23:45 I figure it will be a lot better than nothing 2011-11-11T01:24:22 but I still haven't decided on how to implement combat 2011-11-11T01:25:03 I implemented combat like all enemy ants are coming at me 2011-11-11T01:25:23 So when they're stationary I'm in trouble 2011-11-11T01:25:28 ;_; 2011-11-11T01:25:36 With large pilu-ups 2011-11-11T01:25:40 pile-ups even 2011-11-11T01:27:09 *** ltriant has quit IRC (Quit: Get MacIrssi - http://www.sysctl.co.uk/projects/macirssi/) 2011-11-11T01:28:35 Well stationary is the easiest to implement 2011-11-11T01:31:13 But atleast I didn't time out with 607 ants \o/ 2011-11-11T01:31:50 mmm.. to ditch Java for C++ for this challenge or not 2011-11-11T01:32:56 Well I'm feeling like I need to know another language on top of Python 2011-11-11T01:34:11 *** Antimony has quit IRC (Ping timeout: 256 seconds) 2011-11-11T01:34:47 what is Mu? 2011-11-11T01:35:00 google TrueSkill 2011-11-11T01:35:46 It's the peak of normal distribution 2011-11-11T01:36:06 gah.. these random timeouts..are making me want to pull out C++ 2011-11-11T01:36:17 yyyup 2011-11-11T01:36:18 i random timeout too! 2011-11-11T01:36:33 keep in mind that the time controls are wall clock, too 2011-11-11T01:36:46 so on the official servers you're competing for CPU with everyone 2011-11-11T01:36:58 if you need a lot of cpu you're going to have that much more trouble 2011-11-11T01:37:35 (also the default attempt to avoid timeouts in python uses time.clock() which doesn't work on any system I've tried) 2011-11-11T01:38:41 Oh, you're right 2011-11-11T01:38:55 Just kidding 2011-11-11T01:39:04 time.clock() works fine 2011-11-11T01:39:11 Yeah I sometimes only get 200ms before my doTurn is even called.. like a GC happened or something 2011-11-11T01:39:29 what platform? on every one I've tried time.clock() returns some constant value 2011-11-11T01:39:43 Win XP 2011-11-11T01:39:47 ohh 2011-11-11T01:40:01 well I can assure you that amazon cloud servers are not running on xp :) 2011-11-11T01:40:12 No, they're running Ubuntu 2011-11-11T01:40:55 Linux pulsar 2.6.32-122-rtai #rtai SMP Tue Jul 27 12:44:07 CDT 2010 i686 GNU/Linux 2011-11-11T01:41:03 >>> from time import clock >>> clock() 0.029999999999999999 >>> clock() 0.029999999999999999 >>> clock() 0.029999999999999999 >>> 2011-11-11T01:41:14 O_o 2011-11-11T01:41:26 BenJackson_: yup, constant value here too 2011-11-11T01:41:30 0.06 2011-11-11T01:41:41 also does it on freebsd and some linux servers I tried at work 2011-11-11T01:42:01 Linux 3.1.0-3-ARCH #1 SMP PREEMPT Mon Oct 31 20:19:44 CET 2011 2011-11-11T01:42:03 never occured to me it worked on windows -- I wondered how anyone had happened to use it 2011-11-11T01:42:06 I guess it doesn't work on linux 2011-11-11T01:42:19 >>> time.clock() 155.3739743459015 >>> time.clock() 161.38896619542427 2011-11-11T01:42:27 sure enough, works on win7 2011-11-11T01:42:34 * pairofdice shrugs 2011-11-11T01:43:12 kk.. I don't know C++ all that well.. does it have something like PriorityQueue in Java? 2011-11-11T01:43:22 http://docs.python.org/py3k/library/time.html#time.clock 2011-11-11T01:43:40 *** amstan has quit IRC (Ping timeout: 260 seconds) 2011-11-11T01:44:27 ah, so it just returns the CPU time your process has used I guess 2011-11-11T01:45:03 oh yeah :'( random timeout http://ants.fluxid.pl/replay.29860 2011-11-11T01:45:43 grom358: http://en.cppreference.com/w/cpp/container <- it does 2011-11-11T01:46:09 possibly stupid question: why can't people recover from timing out? 2011-11-11T01:47:02 *** BenJackson_ has quit IRC (Ping timeout: 265 seconds) 2011-11-11T01:47:12 They can avoid time-out 2011-11-11T01:50:04 *** roflmao has quit IRC (Quit: Leaving.) 2011-11-11T01:50:05 are they ignored after the first timeout? 2011-11-11T01:50:23 yeah.. it ignores u after a timeout 2011-11-11T01:50:27 why? 2011-11-11T01:50:44 I added a whole bunch of timeout checks.. but I think a GC is still making me timeout 2011-11-11T01:51:23 u_: your bot is still doing calculations and the other bots need their CPU time 2011-11-11T01:51:45 stopping it and starting it will still leave it stuck in that calculation 2011-11-11T01:52:15 but what is weird.. I added logging and time the lost a turn took.. and the most I ever got was 320ms 2011-11-11T01:52:24 but most of the time its like under 100ms 2011-11-11T01:52:47 the most* 2011-11-11T01:53:08 Heh, I get time-outs on the server when turns take close to 2 seconds locally 2011-11-11T01:53:53 I do BFS for each food limited to depth of 30 tiles.. and then I do a BFS from enemy hill to the closest X ants that is uncapped 2011-11-11T01:55:02 What have you set your turntime as 2011-11-11T01:55:18 okay, another stupid question incoming 2011-11-11T01:55:20 Or are the timeouts on the serve 2011-11-11T01:55:25 is bfs breadth or best first 2011-11-11T01:55:32 breadth 2011-11-11T01:55:33 *** Odyssey has joined #aichallenge 2011-11-11T01:55:52 breadth first 2011-11-11T01:56:10 *** sigh has joined #aichallenge 2011-11-11T01:57:59 *** djr_ has quit IRC (Read error: Connection reset by peer) 2011-11-11T01:58:53 *** djr_ has joined #aichallenge 2011-11-11T02:00:33 *** Odyssey has quit IRC (Quit: Page closed) 2011-11-11T02:02:11 *** grom358 has quit IRC (Quit: Ex-Chat) 2011-11-11T02:03:29 *** dorisabayon has quit IRC (Ping timeout: 252 seconds) 2011-11-11T02:05:27 *** dorisabayon has joined #aichallenge 2011-11-11T02:08:03 *** analyst74 has joined #aichallenge 2011-11-11T02:08:17 *** Garf has joined #aichallenge 2011-11-11T02:10:57 *** BenJackson_ has joined #aichallenge 2011-11-11T02:11:04 arr, web client not so hot 2011-11-11T02:11:41 *** Flort has joined #aichallenge 2011-11-11T02:13:21 *** treeform has joined #aichallenge 2011-11-11T02:14:39 Next game should be within 1851.1 minutes. 2011-11-11T02:14:41 woop 2011-11-11T02:14:53 *** Mooloo has quit IRC (Read error: Connection reset by peer) 2011-11-11T02:15:12 0.2 games per minute, that's a new low 2011-11-11T02:15:40 *** replore has quit IRC (Remote host closed the connection) 2011-11-11T02:16:58 *** ikefruusto has joined #aichallenge 2011-11-11T02:19:34 *** delt0r_ has quit IRC (Ping timeout: 255 seconds) 2011-11-11T02:20:50 *** ikefruusto has quit IRC (Client Quit) 2011-11-11T02:28:58 Next game should be within 494.0 minutes. 2011-11-11T02:32:36 *** delt0r_ has joined #aichallenge 2011-11-11T02:33:29 *** djr_ has quit IRC (Read error: Connection reset by peer) 2011-11-11T02:36:10 *** kilae has joined #aichallenge 2011-11-11T02:37:03 *** Conorach has joined #aichallenge 2011-11-11T02:37:09 *** Accoun has quit IRC (Ping timeout: 258 seconds) 2011-11-11T02:37:24 *** Accoun has joined #aichallenge 2011-11-11T02:38:05 *** tester has joined #aichallenge 2011-11-11T02:38:29 servers seems to be slow 2011-11-11T02:38:38 onlu 0.9 games per minute 2011-11-11T02:38:58 It was down at 0.2, climbing back up 2011-11-11T02:39:04 *** djr_ has joined #aichallenge 2011-11-11T02:39:09 updating workers 2011-11-11T02:39:12 good to hear 2011-11-11T02:40:39 *** tester has quit IRC (Client Quit) 2011-11-11T02:42:28 *** analyst74 has quit IRC (Quit: HydraIRC -> http://www.hydrairc.com <- Nine out of ten l33t h4x0rz prefer it) 2011-11-11T02:44:05 *** jix has quit IRC (Remote host closed the connection) 2011-11-11T02:47:47 janzert: is adding more servers a money problem or a technical problem or an admin time problem? 2011-11-11T02:48:52 money, but we can only accept it through Jeff who has been a bit awol 2011-11-11T02:51:25 *** thestinger has quit IRC (Ping timeout: 240 seconds) 2011-11-11T02:54:37 *** Garf has quit IRC (Quit: Make a new plan, Stan!) 2011-11-11T02:56:10 *** replore_ has joined #aichallenge 2011-11-11T03:07:33 *** datachomper has quit IRC (Quit: Leaving.) 2011-11-11T03:08:53 *** epicmonkey has joined #aichallenge 2011-11-11T03:10:07 *** Garf has joined #aichallenge 2011-11-11T03:10:50 *** LouisMartin has joined #aichallenge 2011-11-11T03:12:14 Is the C# code compiled or interpreted? (mono does both) 2011-11-11T03:13:02 *** djr_ has quit IRC (Read error: Connection reset by peer) 2011-11-11T03:13:36 *** djr_ has joined #aichallenge 2011-11-11T03:14:54 *** epicmonkey has quit IRC (Ping timeout: 240 seconds) 2011-11-11T03:17:20 *** Jak_o_Shadows1 has quit IRC (Remote host closed the connection) 2011-11-11T03:20:03 arg, timeout 3 turns before win :P 2011-11-11T03:20:07 something is crashing, I'm sure :/ 2011-11-11T03:21:09 Fluxid: how do you feel about folks running more than one bot on your server? ie different versions? 2011-11-11T03:22:25 delt0r_: can't see anyone objecting to that. I certainly plan to do so. 2011-11-11T03:22:36 delt0r_: all the more because of the sigma issues on the tcp server 2011-11-11T03:22:38 lots of people already are 2011-11-11T03:22:43 the rankings on fluxid are fucked anyway 2011-11-11T03:22:58 yea i guessed 2011-11-11T03:23:05 just thought i would be polite 2011-11-11T03:23:24 I think I held back until I saw fluxid_vN appear 2011-11-11T03:23:25 My current bot has a few serious bugs... i am surprised it wins at all 2011-11-11T03:23:56 It can't see other ants for example... it thinks its playing all on its own 2011-11-11T03:24:10 and i keeps running away from enemy ant hills 2011-11-11T03:24:24 not to mention it has no concept of its own hills 2011-11-11T03:25:15 Mmm, xathis finished 8th out of ten 2011-11-11T03:26:09 plenty of games are mostly luck 2011-11-11T03:26:25 *** praveen__ has joined #aichallenge 2011-11-11T03:26:30 *** retybok_ has joined #aichallenge 2011-11-11T03:27:22 If you're not xathis 2011-11-11T03:28:13 *** mj41 has joined #aichallenge 2011-11-11T03:29:38 not that you can't construct a game with less of a luck factor 2011-11-11T03:29:56 but in the big 5-6-7-8-way games there can be a lot of chaos 2011-11-11T03:30:02 especially with unbalanced players 2011-11-11T03:30:11 or seemingly insignificant different initial actions 2011-11-11T03:30:59 Of course, of course 2011-11-11T03:31:10 I just want to see xathis fall :) 2011-11-11T03:31:27 (Preferably by my bot) 2011-11-11T03:31:33 =] 2011-11-11T03:32:36 arr I keep thinking "I can wait 30 more minutes to see another game result" 2011-11-11T03:32:43 then the time jumps to 80 minutes... 2011-11-11T03:35:08 *** Blkt has joined #aichallenge 2011-11-11T03:36:17 good morning everyone 2011-11-11T03:40:03 BenJackson_: well, there's always the TCP server... 2011-11-11T03:40:26 but it doesn't give you a very good idea of the ranking you will achieve 2011-11-11T03:52:30 *** g0llum has joined #aichallenge 2011-11-11T04:06:19 *** kiv has quit IRC (Read error: Connection reset by peer) 2011-11-11T04:06:40 *** retybok_ has quit IRC (Ping timeout: 255 seconds) 2011-11-11T04:07:45 *** praveen__ has quit IRC (Quit: Page closed) 2011-11-11T04:09:42 retybok: you get to see how your bot behaves in many different situations however 2011-11-11T04:13:43 delt0r_: i did it myself 2011-11-11T04:15:04 *** aerique has joined #aichallenge 2011-11-11T04:17:06 *** praveen__ has joined #aichallenge 2011-11-11T04:18:16 timedout : am just computing distance from each ant to all food locations. sort based on distances after that. this timedout in tcp server. how? 2011-11-11T04:18:43 *** antimatroid has quit IRC (Ping timeout: 245 seconds) 2011-11-11T04:19:26 am using Java 2011-11-11T04:20:40 *** thestinger has joined #aichallenge 2011-11-11T04:20:54 timedout : am just computing distance from each ant to all food locations. sort based on distances after that. this timedout in tcp server. how? 2011-11-11T04:22:24 it's slow 2011-11-11T04:23:09 I had around 30 ants and 5 food locations. How can that be slow? 2011-11-11T04:23:33 *** retybok_ has joined #aichallenge 2011-11-11T04:23:44 That or your latency to the server is high 2011-11-11T04:24:12 yeah that's what am thinking 2011-11-11T04:25:26 praveen__: it could also be a bug in your code causing infinite loop or excessive iterations 2011-11-11T04:25:48 would it be possible to have the tcpClient correct for lag? 2011-11-11T04:25:56 although it would be an easy exploit I guess :/ 2011-11-11T04:27:09 unlikely all the code I have is two for loops. Am just starting to implement and this is not a good sign. Already thinking to move from Java 2011-11-11T04:27:28 the turn time is 5 seconds, if you have more lag than that you need to test against someone else on earth escape trajectory 2011-11-11T04:27:52 How low can you set turntime locally and not time-out against hunterbot 2011-11-11T04:27:55 Minthos 2011-11-11T04:27:59 :) 2011-11-11T04:29:00 *** retybok_ has quit IRC (Ping timeout: 240 seconds) 2011-11-11T04:33:29 *** retybok_ has joined #aichallenge 2011-11-11T04:33:52 *** Areks has quit IRC (Quit: KVIrc 4.0.2 Insomnia http://www.kvirc.net/) 2011-11-11T04:33:59 *** Areks has joined #aichallenge 2011-11-11T04:34:38 *** Areks has left #aichallenge 2011-11-11T04:35:13 *** djr_ has quit IRC (Ping timeout: 258 seconds) 2011-11-11T04:35:59 Fluxid: right.. thx 2011-11-11T04:38:33 Conorach: its not easy to "correct" for lag 2011-11-11T04:39:07 2 reasons, first it is not constant in time... it changes... and high lag connections to change the most over time 2011-11-11T04:40:27 2ndly, measuring it... is not as easy as it looks. Pings only give you ball park figures. Just look at the ntp protocol to see how to accurately measure lag 2011-11-11T04:40:59 finally what does it really mean? it is easy to add 1000ms to the time out 2011-11-11T04:41:33 praveen_: bugs are bugs... in any lang... do you have much experiance with java 2011-11-11T04:41:55 btw praveen_ I am finding paths with 1000 ants in 50ms with my java bot 2011-11-11T04:43:15 *** jix has joined #aichallenge 2011-11-11T04:44:22 *** retybok_ has quit IRC (Ping timeout: 260 seconds) 2011-11-11T04:47:25 *** Fandekasp has quit IRC (Read error: Operation timed out) 2011-11-11T04:49:40 *** dorisabayon has quit IRC (Ping timeout: 256 seconds) 2011-11-11T05:00:56 *** ikaros has joined #aichallenge 2011-11-11T05:03:47 yeah I guess you're right delt0r, I was thinking of the simple thing there, like: when tcpClient is ran, make some simple ping measurements and add that to the timeout for that bot, just to correct slightly. 2011-11-11T05:04:04 Just adding another second to the timeout internally (while still telling all bots the normal timeout) would help too I guess :) 2011-11-11T05:04:31 or what about timestamping the messages locally by tcpClient ;) 2011-11-11T05:04:33 ah wlel 2011-11-11T05:04:35 well 2011-11-11T05:04:41 People just need decent internet! 2011-11-11T05:18:53 *** dom7b5 has joined #aichallenge 2011-11-11T05:19:54 *** ProperChaos has joined #aichallenge 2011-11-11T05:21:27 *** alc has quit IRC (Quit: Leaving) 2011-11-11T05:21:32 *** b0rder_ has quit IRC (Quit: 离开) 2011-11-11T05:26:26 *** replore_ has quit IRC (Remote host closed the connection) 2011-11-11T05:26:46 *** hihi has joined #aichallenge 2011-11-11T05:27:32 *** retybok_ has joined #aichallenge 2011-11-11T05:30:50 *** praveen__ has quit IRC (Ping timeout: 265 seconds) 2011-11-11T05:42:00 *** retybok_ has quit IRC (Read error: Operation timed out) 2011-11-11T06:00:29 *** MisuArai has joined #aichallenge 2011-11-11T06:16:49 *** grwip has joined #aichallenge 2011-11-11T06:21:08 delt0r: Just noticed your message. Played some more games in tcp server and it did not time out. I hope its not a code issue. 2011-11-11T06:22:18 delt0r: At the same time, many of the top ranked bots are written in Java. So I am going ahead with Java 2011-11-11T06:37:26 what is the tcp server? 2011-11-11T06:41:05 aw nvm found it 2011-11-11T06:53:53 *** Redgis has joined #aichallenge 2011-11-11T06:55:53 *** rintaun has joined #aichallenge 2011-11-11T06:56:27 What exactly are the rules when you want to work in a team? 2011-11-11T06:56:48 We've formed a team after most people already had created an account, although most of us didn't submit anything yet 2011-11-11T06:57:06 praveen_: Languages are rather irrelevant, as you should be able to make a robust sufficiently performant bot in any language you know. 2011-11-11T06:57:39 Can we simply make a new account for the team (even if that gives one or two team members acceess to both the team account and their own account) or how should we handle this? 2011-11-11T06:57:57 *** mleise has joined #aichallenge 2011-11-11T06:58:06 and are individual team members still allowed to submit their own bots independend from the team? 2011-11-11T06:59:39 cool, i can now watch games from a bots perspective live--aka tcp server 2011-11-11T06:59:48 would others be interested? 2011-11-11T07:00:41 praveen_: java is fine... it is pretty fast really, but like all tools.. you need to know how to use it properly 2011-11-11T07:01:11 you can't write java like C and expect performance for example 2011-11-11T07:01:48 its also not magic... but i think for a starting out OO lang --its fine. Hell its fine for production use too 2011-11-11T07:04:33 *** mleyen has joined #aichallenge 2011-11-11T07:05:06 *** twymer has quit IRC (Ping timeout: 258 seconds) 2011-11-11T07:16:47 *** t2027 has quit IRC (Quit: Page closed) 2011-11-11T07:17:11 *** t2027 has joined #aichallenge 2011-11-11T07:51:42 *** mj41 has quit IRC (Ping timeout: 240 seconds) 2011-11-11T07:58:05 *** ikaros_ has joined #aichallenge 2011-11-11T07:59:20 hey, delt0r_, shure we'd all like to have a tcp viz ;) 2011-11-11T08:01:15 spent some hours this morming trying to do that, but my java's pretty rusty, so mixed results here.. 2011-11-11T08:07:18 *** Cybertooth has joined #aichallenge 2011-11-11T08:10:55 *** NoxiaZ^ has joined #aichallenge 2011-11-11T08:11:58 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-11-11T08:14:36 *** ikaros_ has quit IRC (Quit: Ex-Chat) 2011-11-11T08:16:12 ahhh the feeling of a long standing bug finally squashed after a few hours of tracing and fiddling... 2011-11-11T08:16:12 :) 2011-11-11T08:16:51 *** ikaros_ has joined #aichallenge 2011-11-11T08:17:01 Conorach: Causing N new ones. 2011-11-11T08:17:16 *** bergmark has quit IRC (Read error: Operation timed out) 2011-11-11T08:17:43 Ramen to that 2011-11-11T08:17:45 ;) hehe 2011-11-11T08:18:00 well, so far it looks like my bot is working without crashes/bugs/timeout now :) 2011-11-11T08:18:11 but then again, the only thing it does is grabbing food and attacking hills :/ 2011-11-11T08:18:16 pretty mindless still. 2011-11-11T08:18:29 the next bugs come with the next features ;) 2011-11-11T08:19:16 *** TheLinker has joined #aichallenge 2011-11-11T08:20:23 *** ascalino has joined #aichallenge 2011-11-11T08:22:31 hello 2011-11-11T08:22:55 after many turns, my exploration algorithm seems to be tooheavy 2011-11-11T08:23:16 is anybody has an idea too implement a simple "fear of hill" 2011-11-11T08:23:34 to avoid my ants staying to closer to my hill (and blocking new ants) 2011-11-11T08:23:43 *** ztfw has joined #aichallenge 2011-11-11T08:25:46 *** avdg has quit IRC (Quit: Leaving.) 2011-11-11T08:29:18 *** antimatroid has joined #aichallenge 2011-11-11T08:29:47 *** avdg has joined #aichallenge 2011-11-11T08:33:00 *** mj41 has joined #aichallenge 2011-11-11T08:33:00 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Ping timeout: 240 seconds) 2011-11-11T08:33:36 Guys, i got a problem, i tracking where my ants with ID's, and can see that they dont move like they should - Is it possible to track error from "issueorder" - It seems that if you try to move the the same location as a food just spawn, your ant wont move to that location 2011-11-11T08:33:46 *** dwins has joined #aichallenge 2011-11-11T08:34:25 you could just refrain from making an illegal move... 2011-11-11T08:34:48 moving to a place where food is about to spawn is hardly an illegal move 2011-11-11T08:35:00 damn first i didnt know that 2011-11-11T08:35:03 and next. thats just stupid :S 2011-11-11T08:35:35 or are you moving to the food the turn after it spawned? 2011-11-11T08:35:43 moving to a place where food is about to spawn is not illegal, it just blocks the spawn 2011-11-11T08:36:04 Yexo: yeah 2011-11-11T08:36:07 but if food spawns next to your ant you can't move on top of it 2011-11-11T08:36:14 you can check for food collisions before moving 2011-11-11T08:36:50 food is only spawned after the ants have been moved 2011-11-11T08:37:01 *** goffrie_ has joined #aichallenge 2011-11-11T08:37:28 ahh oki, ill check for that - thanks for help =) 2011-11-11T08:37:43 *** goffrie has quit IRC (Ping timeout: 252 seconds) 2011-11-11T08:38:41 worked perfect ! :D 2011-11-11T08:38:47 ^^ 2011-11-11T08:39:10 the other illegal move is on to a hill where there is an ant on the hill and ants in the hive but I will still try to do that one :) 2011-11-11T08:40:00 interesting 2011-11-11T08:40:13 I had so far assumed that would simply kill your ant and the ant on top of the hill since they'd collide 2011-11-11T08:40:46 *** Cybertooth has quit IRC (Ping timeout: 265 seconds) 2011-11-11T08:40:46 *** yoden has joined #aichallenge 2011-11-11T08:41:06 see phases section on http://aichallenge.org/specification.php#Turns-and-Phases 2011-11-11T08:41:41 these phases are done in that order 2011-11-11T08:42:46 don't see anything there about moving to a hill with an ant on it and ants in the hive 2011-11-11T08:43:28 if the exits are blocked, the ants simply don't spawn 2011-11-11T08:43:42 *** TheLinker has quit IRC (Ping timeout: 256 seconds) 2011-11-11T08:44:11 and if there was an ant on top already, but had a collision with an other ant and it didn't move 2011-11-11T08:44:23 you will get a new ant in the same turn 2011-11-11T08:44:30 (because spawning is the last step) 2011-11-11T08:44:37 yes, I see that 2011-11-11T08:44:51 the other illegal move is on to a hill where there is an ant on the hill and ants in the hive but I will still try to do that one :) <- I was referring back to this. How is moving to such a hill illegal? 2011-11-11T08:48:42 You are right I had remembered it being blocking move but it's not. 2011-11-11T08:49:21 *** hacklash has quit IRC (Ping timeout: 260 seconds) 2011-11-11T08:49:43 *** knyppeld1nan has quit IRC (Read error: Connection reset by peer) 2011-11-11T08:55:02 "the other illegal move is on to a hill where there is an ant on the hill and ants in the hive" what? 2011-11-11T08:55:20 I was wrong. misinterpreted the spec. 2011-11-11T08:55:49 Fluxid: is your server down? 2011-11-11T08:56:15 *** rocketpuppy has joined #aichallenge 2011-11-11T08:57:04 *** bergmark has joined #aichallenge 2011-11-11T08:58:16 *** mj41 has quit IRC (Ping timeout: 258 seconds) 2011-11-11T09:01:42 *** ALplus has joined #aichallenge 2011-11-11T09:01:53 *** hihi has quit IRC (Quit: Page closed) 2011-11-11T09:02:58 is it bad if my bot is using 200mb+ of ram? 2011-11-11T09:03:08 no you've got 1GB 2011-11-11T09:03:14 oh that's actually quite nice :] 2011-11-11T09:03:27 Conorach: thats far from 1gig ;-) 2011-11-11T09:03:57 not that I really need that memory, I had some trouble with my deletes, so I removed them for now :/ 2011-11-11T09:04:00 not really proper. 2011-11-11T09:06:07 *** TheLinker has joined #aichallenge 2011-11-11T09:06:38 *** mj41 has joined #aichallenge 2011-11-11T09:11:33 *** ascalino has quit IRC (Quit: Page closed) 2011-11-11T09:16:10 I disabled GC in my bot at one point 2011-11-11T09:16:58 it used up all the machine's memory when I tested it, so I had to kill it 2011-11-11T09:18:23 I am currently not even using 64MB 2011-11-11T09:19:47 @later tell g0llum sure i will have it in the forums some time this afternoon... 2011-11-11T09:19:47 delt0r: Yes master! 2011-11-11T09:20:26 Conorach: what lang are you using? 2011-11-11T09:20:37 C++ :) there are some missing deletes ;) 2011-11-11T09:20:58 I know what is goin wrong but if I try to fix it I get crashes deletin deleted objects 2011-11-11T09:25:24 *** Rinum has joined #aichallenge 2011-11-11T09:25:36 My pathing works flawless now though. only 28s for about 250 ants in ~ 100x100 map. 2011-11-11T09:25:48 28ms 2011-11-11T09:25:49 that is. 2011-11-11T09:26:05 *** MisuArai has quit IRC (Quit: Page closed) 2011-11-11T09:26:19 *** Khaki has joined #aichallenge 2011-11-11T09:26:31 what if you test it on a 200x200 map? 2011-11-11T09:26:46 that should be the max 2011-11-11T09:28:02 which map is good for that? 2011-11-11T09:28:18 hmm 2011-11-11T09:28:21 good question 2011-11-11T09:28:22 *** Migi32 has joined #aichallenge 2011-11-11T09:28:22 i.e which map is the largest? 2011-11-11T09:28:53 I was using maze_13 2011-11-11T09:29:17 tools don't have bigger than 150x150 it seems 2011-11-11T09:29:34 But there's the mapgen 2011-11-11T09:30:10 I was just looking at that. 2011-11-11T09:31:30 but I don't see yet how to set parameters for creating a map :) 2011-11-11T09:33:13 --min_dimensions 2011-11-11T09:34:29 *** bugnuts2 has left #aichallenge ("Leaving") 2011-11-11T09:34:37 *** bugnuts2 has joined #aichallenge 2011-11-11T09:34:41 i test on some 1000x1000 maps! 2011-11-11T09:35:25 The python mapgen doesn't let you create 200x200 maps? Or is it just me 2011-11-11T09:35:37 *** mikewintermute has joined #aichallenge 2011-11-11T09:35:58 Also it gives lots of failures 2011-11-11T09:35:59 Conorach: using valgrind? 2011-11-11T09:36:26 I'm not familiar with that Minthos 2011-11-11T09:36:39 ah debugging tools. 2011-11-11T09:36:41 it's the most valuable debugging tool for C and C++ 2011-11-11T09:36:58 looks pretty interesting. I'll be sure to grab a copy and see how it works. Thanks for pointing that out to me :) 2011-11-11T09:36:59 helps you sort out memory management issues 2011-11-11T09:38:04 hey. Those who use minimax for battles, how does that work? 2011-11-11T09:40:17 you brute force possible moves for your and your opponents' ants and calculate which set of moves is least disadvantageous for you in all scenarios 2011-11-11T09:40:33 then you look for ways to optimize the algorithm so it finishes before your bot times out 2011-11-11T09:40:59 Fluxid: your server seems to be stuck again, I think it needs another kick :P 2011-11-11T09:41:31 Migi32: what do you want to know? how to implement a game tree? how the evaluate states? how to prune? 2011-11-11T09:42:09 Minthos, oh ok. That's actually not minimax. Minimax is about searching a tree. 2011-11-11T09:42:20 *** jasox has joined #aichallenge 2011-11-11T09:42:34 antimatroid, how it's possible to look ahead more than 1 turn 2011-11-11T09:43:21 Migi32: do you mean within the time constraints? or just generally? 2011-11-11T09:43:31 yes, within time constraints 2011-11-11T09:43:38 talk to mcstar 2011-11-11T09:43:53 alpha beta pruning will help once you've generated your first layer 2011-11-11T09:44:01 of course it's possible to brute-force all possible moves, but that's already exponential in number of ants, so doing it for more than 1 turn is just insane 2011-11-11T09:44:20 i think he's also capping it at like 3-4 ants per partition for a game tree 2011-11-11T09:47:38 *** NoxiaZ^ has quit IRC (Ping timeout: 258 seconds) 2011-11-11T09:47:53 do you know which of the top bots uses this kind of battle algorithm? Or what mcstar's username is on the contest server? 2011-11-11T09:48:06 I would like to see it in action :) 2011-11-11T09:48:48 Migi32: the game tree is just how the brute force search is implemented, and it would indeed be impossible to brute force every possible move for all the ants on the map in one round if you did it all at once 2011-11-11T09:49:27 thestinger: kicked 2011-11-11T09:49:39 *** Cowbandit has joined #aichallenge 2011-11-11T09:49:52 like I said you need to optimize it, partitioning the ants into groups is one obvious optimization and alpha-beta pruning is another 2011-11-11T09:50:44 I can maybe see how it's possible, but the way I imagine it now would have some severe downsides and blind spots. For example, capping it at 3-4 ants per partition might make it do very bad things when there's > 4 enemies around 2011-11-11T09:50:58 *** jasox has quit IRC (Quit: Leaving) 2011-11-11T09:50:58 Fluxid: :D works again 2011-11-11T09:51:21 yay 2011-11-11T09:51:55 *** Harpyon has joined #aichallenge 2011-11-11T09:53:53 So it's not 500ms for your bot, it's 500ms for all the bots combined? 2011-11-11T09:54:22 pairofdice: Its 500ms for your bot on the official servers I believe. 2011-11-11T09:54:33 pairofdice, yes, but each of the bots gets its own CPU core 2011-11-11T09:54:37 pairofdice: Remember reading somewhere you get your own core 2011-11-11T09:54:41 that 500ms limit can change btw 2011-11-11T09:54:42 Ahh, okay 2011-11-11T09:54:46 so don't hard code it 2011-11-11T09:57:11 *** NoxiaZ^ has joined #aichallenge 2011-11-11T09:57:31 *** jasox has joined #aichallenge 2011-11-11T10:01:38 *** sigh has quit IRC (Remote host closed the connection) 2011-11-11T10:02:15 I don't understand why my ants still do this "follow the leader" thing 2011-11-11T10:02:56 I penalize the paths taken by other ants, or atleast I think I do 2011-11-11T10:03:42 penalize it more? 2011-11-11T10:09:55 if overpenalizing doesn't work, its probably a bug ;-) 2011-11-11T10:10:00 *** RobotCaleb has joined #aichallenge 2011-11-11T10:10:19 My code has bugs? Never! 2011-11-11T10:10:46 mine stinks of the bugs, but at least I have an idea which parts are stable 2011-11-11T10:11:37 *** datachomper has joined #aichallenge 2011-11-11T10:12:40 my bot has lots of bugs too. Ants, to be specific. ;) 2011-11-11T10:14:07 my ants are stupid in my model 2011-11-11T10:14:31 (so they don't have codes, except a few states) 2011-11-11T10:14:57 though I wish I could code faster and better 2011-11-11T10:15:42 I'm pretty happy with how my ants wait for reinforcements now 2011-11-11T10:16:53 The version on the server gets stuck in death stares even with 20 of my guys against 1 of theirs 2011-11-11T10:17:30 *** cbad has quit IRC (Read error: Connection reset by peer) 2011-11-11T10:17:44 I implemented pathfinding with higher costs for some tiles 2011-11-11T10:17:55 *** cbad has joined #aichallenge 2011-11-11T10:17:56 *** cbad has joined #aichallenge 2011-11-11T10:18:23 speaking of bugs: the website has a bug. If you're on a page with a visualizer, you can't type certain letters (like "i") in the user search box 2011-11-11T10:18:54 *** delt0r_ has quit IRC (Read error: Operation timed out) 2011-11-11T10:19:29 yeah, I've added a bug report 2011-11-11T10:19:55 but no response so far (not that I really care) 2011-11-11T10:20:31 https://github.com/aichallenge/aichallenge/issues/360 2011-11-11T10:21:13 but maybe I'm just bad in explaining things in english 2011-11-11T10:22:31 no, it's perfectly understandable. I think the admins are just busy with other things atm 2011-11-11T10:22:54 it's only a small inconvenience anyway, not a big issue 2011-11-11T10:23:14 *** Antimony has joined #aichallenge 2011-11-11T10:23:26 *** ALplus has quit IRC (Ping timeout: 256 seconds) 2011-11-11T10:25:03 *** NoxiaZ^ has quit IRC (Ping timeout: 248 seconds) 2011-11-11T10:26:13 oh, I didn't notice that the tcp server is again working 2011-11-11T10:27:32 *** hacklash has joined #aichallenge 2011-11-11T10:32:32 *** treeform has quit IRC (Remote host closed the connection) 2011-11-11T10:32:47 *** kilae has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111104165243]) 2011-11-11T10:32:54 *** delt0r_ has joined #aichallenge 2011-11-11T10:34:57 *** aerique has quit IRC (Quit: ...) 2011-11-11T10:40:22 pairofdice: You can get "follow the leader" behavour quite easily 2011-11-11T10:40:36 for example "go to the closest peice of food" 2011-11-11T10:41:09 if there are 4 ants where its the same piece of food... you get follow the leader 2011-11-11T10:44:50 *** yoR has joined #aichallenge 2011-11-11T10:45:47 * avdg thinks his bot is breaking everything atm 2011-11-11T10:47:48 *** conor_f has joined #aichallenge 2011-11-11T10:50:14 *** amstan has joined #aichallenge 2011-11-11T10:50:14 *** ChanServ sets mode: +o amstan 2011-11-11T10:51:27 *** yoR has quit IRC (Quit: Page closed) 2011-11-11T10:52:39 *** moondust has quit IRC (Remote host closed the connection) 2011-11-11T10:56:33 *** sheerun has joined #aichallenge 2011-11-11T10:56:40 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-11T10:59:10 *** Redgis has quit IRC (Quit: ... mains libres) 2011-11-11T11:00:32 *** Blkt has quit IRC (Remote host closed the connection) 2011-11-11T11:00:49 *** retybok_ has joined #aichallenge 2011-11-11T11:00:55 lol, my bot actually won a game x.x http://ants.fluxid.pl/replay.30306 2011-11-11T11:01:42 nice one. 2011-11-11T11:02:02 my still shouldn't win 2011-11-11T11:04:00 *** Antimony has joined #aichallenge 2011-11-11T11:07:07 *** JorgeB has quit IRC (Read error: Operation timed out) 2011-11-11T11:07:32 *** regularguy has joined #aichallenge 2011-11-11T11:07:54 lol... looks like my bot has a favorite map 2011-11-11T11:09:37 *** mcstar has joined #aichallenge 2011-11-11T11:10:01 *** ikaros_ has quit IRC (Quit: Ex-Chat) 2011-11-11T11:12:06 *** retybok_ has quit IRC (Ping timeout: 276 seconds) 2011-11-11T11:12:31 *** conor_f has quit IRC (Ping timeout: 248 seconds) 2011-11-11T11:23:46 just made my "live viewer" available 2011-11-11T11:23:56 on the forums 2011-11-11T11:24:15 http://aichallenge.org/forums/viewtopic.php?f=25&t=1861 2011-11-11T11:24:21 *** knott has joined #aichallenge 2011-11-11T11:24:23 * delt0r goes to the beer hour 2011-11-11T11:24:38 *** conor_f has joined #aichallenge 2011-11-11T11:24:49 :-) 2011-11-11T11:24:56 *** djr_ has joined #aichallenge 2011-11-11T11:25:02 Hello. Why Java code compiles only with ASCII encoding in submisson? 2011-11-11T11:25:33 *** djr_ has quit IRC (Read error: Connection reset by peer) 2011-11-11T11:26:01 *** ChrisH_ has joined #aichallenge 2011-11-11T11:26:17 delt0r: nice! 2011-11-11T11:27:17 @later tell McLeopold The hover text for game times in game lists that tells you how long ago a game was seems to be off by about an hour. 2011-11-11T11:27:17 ChrisH_: I think that worked... 2011-11-11T11:27:33 delt0r, looks awesome. I'm going to try it when version 2 of my bot is done. 2011-11-11T11:28:40 delt0r: woah, cool! 2011-11-11T11:28:50 delt0r: tell me when the 9001 edition comes along 2011-11-11T11:32:39 *** JorgeB has joined #aichallenge 2011-11-11T11:32:56 so its a "hud" display? 2011-11-11T11:33:22 anyway, my comment of the day is: god bless valgrind 2011-11-11T11:37:22 *** TheLinker has quit IRC (Quit: Bye) 2011-11-11T11:38:06 *** jasox has quit IRC (Ping timeout: 240 seconds) 2011-11-11T11:42:49 *** praveen has joined #aichallenge 2011-11-11T11:44:29 *** ChrisH_ has quit IRC (Quit: Page closed) 2011-11-11T11:49:10 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-11T11:50:14 *** Rinum has quit IRC (Ping timeout: 265 seconds) 2011-11-11T11:56:47 *** amstan has quit IRC (Ping timeout: 248 seconds) 2011-11-11T11:57:43 *** mj41 has quit IRC (Read error: Operation timed out) 2011-11-11T11:57:49 *** Losealot has joined #aichallenge 2011-11-11T12:00:26 *** Antimony has joined #aichallenge 2011-11-11T12:02:28 *** sheerun has quit IRC (Quit: Page closed) 2011-11-11T12:04:43 *** praveen has quit IRC (Quit: Page closed) 2011-11-11T12:08:26 *** Antimony_ has joined #aichallenge 2011-11-11T12:08:29 *** g0llum has joined #aichallenge 2011-11-11T12:09:18 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-11-11T12:09:21 *** Antimony_ is now known as Antimony 2011-11-11T12:16:38 *** Hexren has joined #aichallenge 2011-11-11T12:16:43 *** Hexren has left #aichallenge 2011-11-11T12:21:08 *** Antimony_ has joined #aichallenge 2011-11-11T12:22:45 *** Kingpin13 has quit IRC (Quit: quit) 2011-11-11T12:23:32 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-11T12:23:50 *** Antimony has joined #aichallenge 2011-11-11T12:25:56 lol 2011-11-11T12:26:02 next game in -0.3 mins :P 2011-11-11T12:26:21 refresh and now it's 15 >_< 2011-11-11T12:26:34 :-) 2011-11-11T12:27:07 *** Antimony_ has quit IRC (Ping timeout: 260 seconds) 2011-11-11T12:28:49 http://ants.fluxid.pl/replay.30360 <- look how only 1 ant captures the flag after a big battle :-) 2011-11-11T12:29:27 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-11T12:30:03 *** Antimony has joined #aichallenge 2011-11-11T12:30:18 or almost 1 2011-11-11T12:34:43 avdg: yeah, that's pretty cool :) I like how so many ants died 1 square away from the hill 2011-11-11T12:35:15 *** Kingpin13 has joined #aichallenge 2011-11-11T12:36:40 meh, now I'm realizing my circle algorithm isn't fitting with the game implementation 2011-11-11T12:42:06 does anyone know if it really is supposed to be several hours between games? 2011-11-11T12:42:47 Losealot: use the tcp server if you want to play lots of games 2011-11-11T12:42:53 I had to wait around 6 hours between every game, but that wait time is raised a little bit 2011-11-11T12:46:09 *** datachomper has left #aichallenge 2011-11-11T12:49:08 thanks thestinger 2011-11-11T12:49:28 @tcp 2011-11-11T12:49:29 avdg: tcp could be http://ants.fluxid.pl/howto. 2011-11-11T12:49:52 @rainbow [tcp] 2011-11-11T12:49:53 thestinger: tcp could be http://ants.fluxid.pl/howto. 2011-11-11T12:49:58 better :P 2011-11-11T12:50:20 whoa. @rainbow [seen thestinger] 2011-11-11T12:50:22 er. 2011-11-11T12:50:26 @rainbow [seen thestinger] 2011-11-11T12:50:26 a1k0n: thestinger was last seen in #aichallenge 27 seconds ago: better :P 2011-11-11T12:50:34 haha that's just great 2011-11-11T12:51:01 greetings gentlemen 2011-11-11T12:51:07 :) 2011-11-11T12:51:20 *** fufu94 has joined #aichallenge 2011-11-11T12:51:41 hi 2011-11-11T12:51:46 who can tell me if C# is compiled or interpreted? (mono does both) 2011-11-11T12:52:35 somebody speak french here pls ? 2011-11-11T12:53:01 yep 2011-11-11T12:53:21 tu parle français ? 2011-11-11T12:53:22 fufu94: I know a tiny amount of french so if you're looking for an explanation of the game, sorry from me :/ 2011-11-11T12:53:30 fufu94: I do 2011-11-11T12:53:33 oui, je suis français... 2011-11-11T12:53:38 lol :P 2011-11-11T12:53:55 grwip: flawless victory :P 2011-11-11T12:54:11 ok cool ^^ 2011-11-11T12:54:26 not always proud but french... 2011-11-11T12:54:29 esque tu pourrai me dire , coment lancer le pack de test ? 2011-11-11T12:54:41 pasque ja bien lut le tuto et tout 2011-11-11T12:54:53 je n'aime pas parle en français 2011-11-11T12:54:54 :p 2011-11-11T12:55:03 mais chez moi le .cmd se lance pas =\ 2011-11-11T12:55:22 dsl mais moi je ne sais pas bien parler en anglais ^^ 2011-11-11T12:55:37 :-) 2011-11-11T12:55:56 fufu94: private chat (fenêtre de chat privée qui doit s'être puverte chez toi) 2011-11-11T12:56:02 gawd.. every try makes my exploration strategy worse 2011-11-11T12:56:06 i dont get this 2011-11-11T12:56:41 does chez moi/toi not mean my/your house? 2011-11-11T12:56:52 i was just gonna say that 2011-11-11T12:57:06 yep 2011-11-11T12:57:07 fufu then said but my house the .cmd doesn't throw? 2011-11-11T12:57:15 lol, my french is soo poor :P 2011-11-11T12:57:22 but meaning more 'on my machine'... 2011-11-11T12:57:30 i guess ordinateurs are houses now 2011-11-11T12:57:31 ah, ok 2011-11-11T12:57:35 :P 2011-11-11T12:57:50 a1k0n: some of us live in computers, don't judge :) 2011-11-11T12:57:52 lancer == throw? or is it another phrase? 2011-11-11T12:58:02 haha ikaros that is the same for me :/ 2011-11-11T12:58:12 yes it is, meaning 'run' 2011-11-11T12:58:17 it's like I got a decent bot, then 2 hours later it is only doing shitty things like bang into walks and suicide again :P 2011-11-11T12:58:25 lancer is like launch :] 2011-11-11T12:58:28 Argh... 500 ants in the hive, guess that version has a few issues 2011-11-11T12:58:32 i had a simple exploration like 2 weeks ago 2011-11-11T12:58:48 just moving ants in the same direction until they bounc against something 2011-11-11T12:58:49 *** knott has quit IRC (Quit: Page closed) 2011-11-11T12:59:04 in a computer? I would prefer it above my cold room I'm currently at :-) 2011-11-11T12:59:10 and then changed direction 2011-11-11T12:59:12 sad its a laptop 2011-11-11T12:59:16 *** Antimony has quit IRC (Ping timeout: 256 seconds) 2011-11-11T12:59:22 this works nice for maps which arent mazes 2011-11-11T12:59:42 then i tried some diffusion stuff.. other strategies and they all sucked 2011-11-11T12:59:52 grwip: yeah, I just know one word for everything I know in french. courir is run to me :P 2011-11-11T12:59:56 and i thought i could finally step to combat.. god 2011-11-11T13:00:01 look at me learning french lol 2011-11-11T13:00:11 hard language... ;) 2011-11-11T13:00:20 :P I like it tbh 2011-11-11T13:00:25 would like to improve though 2011-11-11T13:00:30 i have to say though, if you're going to open windows you might as well stick with the metaphor 2011-11-11T13:02:20 *** peyton has quit IRC (Quit: peyton) 2011-11-11T13:03:11 2111 locs 2011-11-11T13:03:13 *** praveen_ has quit IRC (Ping timeout: 265 seconds) 2011-11-11T13:05:00 huh 2011-11-11T13:05:02 great 2011-11-11T13:05:17 i made a tiny change and my bot eats up 6gb memory like candy 2011-11-11T13:08:05 hey, who's the definitive authority on the game engine that runs the contest? 2011-11-11T13:08:23 *** pedrosorio has joined #aichallenge 2011-11-11T13:08:32 *** Losealot has quit IRC (Ping timeout: 265 seconds) 2011-11-11T13:12:31 *** fufu94 has quit IRC (Quit: Page closed) 2011-11-11T13:14:02 *** amstan has joined #aichallenge 2011-11-11T13:14:02 *** ChanServ sets mode: +o amstan 2011-11-11T13:14:03 janzert or antimatroid? 2011-11-11T13:14:32 hey amstan, who's the definitive authority on the game engine that runs the contest and how can I direct him my questions? 2011-11-11T13:14:49 define authority 2011-11-11T13:15:18 if you're asking who knows a lot about.. kinda of question 2011-11-11T13:15:18 who knows best about it. 2011-11-11T13:15:26 mcleopold and janzert would be your best bet 2011-11-11T13:15:31 about the compilation of the codes in different language 2011-11-11T13:15:38 i might be able to answer that too 2011-11-11T13:15:53 it's a good idea to ask the question, instead of ask to ask 2011-11-11T13:16:21 ok. The way the code is interpreted or compiled makes a huuuge difference in what we can or can't do in C# 2011-11-11T13:16:34 amstan: I did, but nobody answered 2011-11-11T13:17:05 LouisMartin: ok, what about it? you think we're doing something that's slow for c#? 2011-11-11T13:17:39 amstan: first thing to know is: mono can both compiled and interpret C#. Do you compiled it or do you interpret it in the contest? 2011-11-11T13:18:11 https://github.com/aichallenge/aichallenge/blob/epsilon/worker/compiler.py has all the info on how stuff gets compiled 2011-11-11T13:18:19 it is a clr binary 2011-11-11T13:18:23 btw, I guess since multi-threading is not allowed you can't use timers in C# right? 2011-11-11T13:18:26 "C#" : [["gmcs", "-warn:0", "-out:%s.exe" % BOT]], 2011-11-11T13:18:31 it is not a source that gets interpreted 2011-11-11T13:18:36 grwip: depends on the type of thread 2011-11-11T13:18:53 amstan: I'll take a look, thanks. 2011-11-11T13:19:10 mcstar: ok, that a first great news. :P 2011-11-11T13:19:25 LouisMartin: there's exes being produce, so i assume they're compiled 2011-11-11T13:19:36 oh ok, so some kind of multi threadings are allowed? 2011-11-11T13:19:49 grwip: if it's a green thread yes 2011-11-11T13:20:03 amstan: ok thx for the precision 2011-11-11T13:20:15 grwip: you want to terminate one thread while it is running by a timer thread? 2011-11-11T13:20:17 The optimisation switch makes a world of difference in C#, so that's my next question (which may be answered in the url amstan gave me) 2011-11-11T13:20:49 LouisMartin: as long as nothing breaks in people's submissions, we usually accept any changes to the flags 2011-11-11T13:21:05 mcstar: I thought about it yes, to avoid timeouts, but I code my bot in java and I don't know how timers works there, but I was wondering for C#... 2011-11-11T13:21:28 grwip: green threads are equivalent to doing it manually 2011-11-11T13:21:33 grwip: my intuition says you need os level threads for that, so probably wont work 2011-11-11T13:21:48 maybe on java it will, cause it is enabled for java to spawn multiple threads 2011-11-11T13:22:11 amstan: that's awesome. you should charge participation fees ;) btw, how did you guys end up putting this contest up together? 2011-11-11T13:23:05 LouisMartin: it started as a small contest(10 participants) inside the computer science club at university of waterloo 2011-11-11T13:23:18 mcstar: Interesting, and if you submit a bot that doesn't respect the multi-threading clause, will you be noticed after compilation on the server? 2011-11-11T13:23:32 at runtime 2011-11-11T13:23:33 grwip: thread creation will fail 2011-11-11T13:23:57 ok, so it'll just not work and may be kind of hard to detect... 2011-11-11T13:24:06 amstan: cool 2011-11-11T13:24:22 LouisMartin: it's more awesome if it's free 2011-11-11T13:25:07 Yeah, this is pretty awesome 2011-11-11T13:25:11 amstan: that's right but is there some kind of paypal donate stuff somewhere where one can show support? 2011-11-11T13:25:25 (not really show, just support) 2011-11-11T13:25:26 grwip: yeah, we might get that in the next month or so 2011-11-11T13:25:59 this would be a good idea, this challenge is so nice I bet some people would give you something (I guess I'll do!) 2011-11-11T13:26:00 *** treeform has joined #aichallenge 2011-11-11T13:26:02 i say the winner should be fined 2011-11-11T13:26:08 amstan: of course! It was my way to show how much I appreciate the contest and that I know how much effort you guys put into making it that awesome 2011-11-11T13:26:15 mcstar: or what? lol 2011-11-11T13:26:33 amstan: he will cost us many hours of work 2011-11-11T13:27:05 mcstar: yes, and what if he doesn't want to pay? 2011-11-11T13:27:20 amstan: kidding 2011-11-11T13:27:33 let the bastard have all the fame he can get 2011-11-11T13:27:36 mcstar: not like we can claim he didn't win anymore 2011-11-11T13:27:41 ...in jail! 2011-11-11T13:28:32 old fashion stake? 2011-11-11T13:28:41 yes, hire assasing 2011-11-11T13:28:45 assasins* 2011-11-11T13:28:54 assassins* 2011-11-11T13:29:21 Hashishin 2011-11-11T13:29:55 interesting that assassination ends in nation 2011-11-11T13:30:18 hmm, I'll go meditate on that one, bbl 2011-11-11T13:30:40 what is the stderr only pipe operator? 2011-11-11T13:30:48 | stdout, |& both 2011-11-11T13:31:28 mcstar: good to know |& 2011-11-11T13:31:45 yeah, i knew the other one too 2011-11-11T13:31:48 you can swap the file descriptors iirc 2011-11-11T13:31:50 but the manual is huge 2011-11-11T13:32:09 isn't it 2> ? 2011-11-11T13:32:16 Migi32: that goes to a file 2011-11-11T13:32:24 ah yeah, true 2011-11-11T13:32:48 k, i was wrong 2011-11-11T13:32:53 |& is stderr only 2011-11-11T13:33:00 .o) 2011-11-11T13:33:02 *** xathis has joined #aichallenge 2011-11-11T13:33:11 then the engine speaks on stderr: 2011-11-11T13:33:12 is that some sort of a messed up smiley or what? 2011-11-11T13:33:13 ? 2011-11-11T13:33:15 2>&1 2011-11-11T13:33:57 hmm you want to pipe stderr into stdin of another process? 2011-11-11T13:34:40 ok, i think less doesnt respect me 2011-11-11T13:34:44 what i did: 2011-11-11T13:34:53 ./playonesomething |& less 2011-11-11T13:34:58 to capture my bots stderr 2011-11-11T13:35:08 but ./playgame.py's stdout shows up too 2011-11-11T13:35:23 I think |& does both 2011-11-11T13:35:33 thestinger: no 2011-11-11T13:35:39 i looked up in man bash 2011-11-11T13:35:42 says stderr only 2011-11-11T13:35:51 but I tried it and it seems to do both :P 2011-11-11T13:36:04 If |& is used, the standard error of command is 2011-11-11T13:36:14 it is shorthand for 2>&1 | 2011-11-11T13:36:30 thestinger: yep, but i think that is less's fault 2011-11-11T13:36:50 I think 2>&1 will still send stdout 2011-11-11T13:37:07 yes, that makes stderr go into stdout, and then both get piped 2011-11-11T13:37:10 https://secure.wikimedia.org/wikipedia/en/wiki/Named_pipe <- just use one of those 2011-11-11T13:37:22 oh 2011-11-11T13:37:24 spectacular idea. 2011-11-11T13:37:30 I think you can do a fancy swap of stdout and stderr with like 4 redirections but fuck that 2011-11-11T13:37:57 *** foRei has joined #aichallenge 2011-11-11T13:38:13 fuck that 2011-11-11T13:39:12 maybe `command 2>&1 >/dev/null' will work 2011-11-11T13:42:43 ./cpp.sh > /dev/null | less 2011-11-11T13:42:48 this is my solution 2011-11-11T13:43:17 *** Cowbandit has quit IRC (Quit: Page closed) 2011-11-11T13:44:34 amstan: I just read that about the C# optimisation, are you going to add this (-O=all) to the compiler flags? 2011-11-11T13:44:56 ProperChaos: how much does that slow down compilation though 2011-11-11T13:44:59 there is a tradeoff 2011-11-11T13:45:13 good point, I have no idea 2011-11-11T13:45:14 all optimization doesn't necessarly mean faster 2011-11-11T13:45:27 might even slow down some code 2011-11-11T13:46:38 yeah true, although I think for most people it will help, for me it's almost a 50% improvement 2011-11-11T13:46:59 but I get your point 2011-11-11T13:47:24 *** Antimony has joined #aichallenge 2011-11-11T13:48:22 *** NoxiaZ^ has joined #aichallenge 2011-11-11T13:50:19 how much time do the servers spend compiling, compared to playing games? 2011-11-11T13:50:36 nothing at all? 2011-11-11T13:50:54 then is slower compilation really a concern?µ 2011-11-11T13:51:12 i think it is not distributed 2011-11-11T13:52:24 still, C# compilation takes 5 seconds with optimisations at most 2011-11-11T13:54:16 *** fufu94 has joined #aichallenge 2011-11-11T13:54:50 5 seconds was too much for me, i split my bot into 20 files 2011-11-11T13:55:09 Unless people do generate large complicated data sets at compile time, it should be fine 2011-11-11T13:55:50 But we had only a couple of participants when I last thought about it. *g* 2011-11-11T13:56:13 soon there will be a couple of 10k 2011-11-11T13:56:23 mcstar: I've got 16 files, takes 1-2 seconds locally for me 2011-11-11T13:56:39 how many locs? 2011-11-11T13:56:41 c++? 2011-11-11T13:56:47 mine is c++11 2011-11-11T13:56:56 and uses stl massively 2011-11-11T13:56:59 and lambdas 2011-11-11T13:57:21 and i have my own templates too 2011-11-11T13:57:26 *** Khaki has quit IRC (Quit: Leaving) 2011-11-11T13:57:26 *** ajf|offline is now known as ajf 2011-11-11T13:57:58 actually, i dont change much my templated classes, i pre-compiled them int a precompiled header 2011-11-11T13:57:59 no c#, which is a lot faster at compiling 2011-11-11T13:58:31 *** mleyen has quit IRC (Quit: ~ Trillian Astra - www.trillian.im ~) 2011-11-11T13:59:12 mcstar: make sure you don't use any c++11 features that are only available in gcc 4.6 and up. I had to rewrite all my fancy for (auto& x : y) loops to c++98 for-loops 2011-11-11T13:59:33 yeah, cant use range based fors 2011-11-11T13:59:59 Wow! Blue's ants are really scared. http://aichallenge.org/visualizer.php?game=62871 2011-11-11T14:01:35 *** rofer is now known as Rofer 2011-11-11T14:01:36 haha 2011-11-11T14:02:01 purple's ants are very lazy 2011-11-11T14:02:07 *** Rofer is now known as rofer 2011-11-11T14:02:29 blue prefers to envelop rather than kill. he's very humane. 2011-11-11T14:02:48 that match demonstrates the timeout defense really well 2011-11-11T14:03:06 i can't figure out how blue keeps getting more ants. i guess food is spawning on top of his massive pile? 2011-11-11T14:04:41 my next game has been negative minutes for the last few hourst :( 2011-11-11T14:05:32 well he did have a few ants out gathering 2011-11-11T14:05:32 conor_f: have you checked that your submission isn't deactivated? 2011-11-11T14:05:35 on your profile 2011-11-11T14:05:53 http://aichallenge.org/visualizer.php?game=83191&user=8217 2011-11-11T14:06:01 purples defense sort of thing is cool 2011-11-11T14:06:05 and Migi32 tyvm :P 2011-11-11T14:06:17 oh, no problem :) 2011-11-11T14:07:12 pretty fun behavior :) 2011-11-11T14:07:19 yeah, effective too 2011-11-11T14:07:30 pity it's not my one :P 2011-11-11T14:07:39 What was purple doing? It didn't seem particularly special to me 2011-11-11T14:08:11 stays static until a number of ants are in the hive 2011-11-11T14:08:30 then moves a bit, then stays static again except with more ants gathering food 2011-11-11T14:08:52 Encapsulation complete! 2011-11-11T14:09:38 I like my defense strategy the best 2011-11-11T14:09:57 Each turn, calculate the closest possible enemy position and force 1/5th of ants to stay at lesat that close 2011-11-11T14:10:04 *** mikewintermute has quit IRC (Quit: mikewintermute) 2011-11-11T14:10:30 It works well as long that you don't get a concerted attack from someone with a numerical advantage 2011-11-11T14:10:39 heh, cool 2011-11-11T14:11:03 what do you do when your number of ants is not divisible with 5? 2011-11-11T14:11:08 but the main advantage is that guards can gather food too as long as no ants are near 2011-11-11T14:11:18 mcstar: just cast it :P 2011-11-11T14:11:22 http://aichallenge.org/visualizer.php?game=85966&user=8217 2011-11-11T14:11:22 it's approximately 1/5th. The actual formula is slightly more complex 2011-11-11T14:11:23 *** mikewintermute has joined #aichallenge 2011-11-11T14:11:36 *** mikewintermute has quit IRC (Client Quit) 2011-11-11T14:11:38 talk about a shitstorm of ants at the end :O 2011-11-11T14:12:06 bleh 2011-11-11T14:12:14 was supposed to be somehwere 5 mins ago 2011-11-11T14:12:15 ttyl 2011-11-11T14:14:13 *** lavalamp has joined #aichallenge 2011-11-11T14:15:39 *** BenJackson has quit IRC (Quit: Page closed) 2011-11-11T14:15:58 *** dvladim has joined #aichallenge 2011-11-11T14:16:44 *** conor_f has quit IRC (Ping timeout: 260 seconds) 2011-11-11T14:27:24 amstan: About the C# compilation, there's no --optimize=all done and after a few test, I observe quit a dramatic difference. 2011-11-11T14:29:07 amstan: *quite. 500+ ms timeouts are tamed into gentle ~50ms 2011-11-11T14:29:35 LouisMartin: what about compile time? 2011-11-11T14:30:25 *** retybok_ has joined #aichallenge 2011-11-11T14:30:25 amstan: hard to say as it's almost instantly in both cases and I have probably 2 times more codes than 80% of the bot out there 2011-11-11T14:30:33 *** Hexren1 has joined #aichallenge 2011-11-11T14:33:46 amstan: One could argue that with such results I could use a few manual optimizations, but C#/Mono (CLR) isn't exactly C++ and some things can only be optimized by the compiler. Garbage Collection is terrible thing when ms makes a difference. You can't optimized that well in managed language and you need all the help your compiler offer. 2011-11-11T14:34:07 *** Antimony has quit IRC (Ping timeout: 248 seconds) 2011-11-11T14:34:33 wow; my english is disgusting today. sorry about that. 2011-11-11T14:34:51 LouisMartin: ok, here's my suggestion, start a discussion about this either on github(pull request) or forums(perhaps a poll in the language section) 2011-11-11T14:35:01 if most people agree with you, we'll do it 2011-11-11T14:35:34 you mean C# people, right? 2011-11-11T14:35:37 yes 2011-11-11T14:36:02 okay 2011-11-11T14:36:38 I'll do 2011-11-11T14:37:33 *** Antimony has joined #aichallenge 2011-11-11T14:37:53 *** retybok_ has quit IRC (Ping timeout: 245 seconds) 2011-11-11T14:40:02 *** xathis has quit IRC (Read error: Connection reset by peer) 2011-11-11T14:48:50 *** amstan has quit IRC (Ping timeout: 260 seconds) 2011-11-11T14:49:24 *** Dabrion has joined #aichallenge 2011-11-11T14:49:28 *** LouisMartin has quit IRC (Quit: Page closed) 2011-11-11T14:50:30 *** choas has joined #aichallenge 2011-11-11T14:51:05 *** Antimony_ has joined #aichallenge 2011-11-11T14:51:28 *** Antimony has quit IRC (Ping timeout: 256 seconds) 2011-11-11T14:51:31 *** Antimony_ is now known as Antimony 2011-11-11T14:56:22 *** Accoun has quit IRC () 2011-11-11T15:04:55 *** dwins has quit IRC (Ping timeout: 252 seconds) 2011-11-11T15:07:30 *** dwins has joined #aichallenge 2011-11-11T15:11:44 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-11-11T15:13:03 *** mleise has quit IRC (Ping timeout: 248 seconds) 2011-11-11T15:13:40 *** Zaphus has joined #aichallenge 2011-11-11T15:14:06 *** pedrosorio has quit IRC (Quit: Page closed) 2011-11-11T15:16:17 *** Accoun has joined #aichallenge 2011-11-11T15:17:43 Is there something up with the forums ? suddenly it's prompting me for my login, but won't accept it - and the forgotten password says my username/email combo doesnt exist ?! 2011-11-11T15:18:48 hmm 2011-11-11T15:18:50 i think it's not the same as your bot account there 2011-11-11T15:18:59 mine works fine 2011-11-11T15:19:29 but the forgot password should still work - hmmm 2011-11-11T15:19:51 lets hope it works :-) 2011-11-11T15:21:17 ah, I see what happened... Chrome Sync has messed things up for me :-/ 2011-11-11T15:22:01 ah, did it swap passwords between logins? 2011-11-11T15:22:05 it happened to me 2011-11-11T15:23:09 yeah, something like that - actually switched the username to another I regularly use 2011-11-11T15:36:40 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-11T15:37:34 *** ALplus has joined #aichallenge 2011-11-11T15:41:49 *** conor_f has joined #aichallenge 2011-11-11T15:42:14 *** rocketpuppy has quit IRC (Ping timeout: 265 seconds) 2011-11-11T15:52:12 *** retybok_ has joined #aichallenge 2011-11-11T15:53:18 *** dvladim has quit IRC (Ping timeout: 240 seconds) 2011-11-11T15:53:23 *** Fandekasp has joined #aichallenge 2011-11-11T15:53:32 *** amstan has joined #aichallenge 2011-11-11T15:53:32 *** ChanServ sets mode: +o amstan 2011-11-11T15:53:33 *** dorisabayon has joined #aichallenge 2011-11-11T15:54:22 *** pairofdice has quit IRC (Quit: Just keep breathing) 2011-11-11T16:05:15 *** retybok_ has quit IRC (Ping timeout: 276 seconds) 2011-11-11T16:12:27 wow, my bot got given 4 games in a single hour last night... thats as many as I normally get in an entire day! 2011-11-11T16:17:44 :-) 2011-11-11T16:23:08 *** Anilm3 has joined #aichallenge 2011-11-11T16:23:14 Next Game: 942 players are ahead. 2011-11-11T16:23:42 Planet Wars wasn't this slow 2011-11-11T16:24:36 *** roflmao has joined #aichallenge 2011-11-11T16:24:56 nope, it wasn't 2011-11-11T16:24:59 *** retybok_ has joined #aichallenge 2011-11-11T16:26:44 hmmmmmm 2011-11-11T16:27:33 planet wars had fewer entries, but the difference isn't enough to explain the slowness, I think 2011-11-11T16:28:09 it's because there's up to 8times more players per game 2011-11-11T16:28:38 *** fufu94 has quit IRC (Ping timeout: 265 seconds) 2011-11-11T16:28:57 That would decrease the number of games, but it shouldn't decrease players per minute, no? 2011-11-11T16:29:39 I wonder how much harder this problem is to solve 2011-11-11T16:29:48 There's more standoffs in this, so the games go longer. 2011-11-11T16:29:57 *** roflmao has left #aichallenge 2011-11-11T16:30:22 That probably explains a lot of it. 2011-11-11T16:30:50 I dont think there were many PW games where the two sides stood across from each other and refused to make the first move :-) 2011-11-11T16:31:35 is this better or worse then PW? 2011-11-11T16:31:41 I wasn't around for it 2011-11-11T16:32:18 plus pathfinding, PW had none - thats a huge amount of time per turn... call it 0.1s per player per turn, thats around 0.5s per turn across the game, for 1000 turns... easily 5 minutes per game eaten by that (if the game lasts) - could be much higher 2011-11-11T16:33:16 http://aichallenge.org/game_settings.php is this what is really running on the servers and what i should use for local testing ? 2011-11-11T16:33:25 *** pedrosorio has joined #aichallenge 2011-11-11T16:33:41 I think the bottom half of bots don't do any pathfinding ;) 2011-11-11T16:34:36 I just do basic, basic pathfinding and I'm nearly in the top 1000 :P 2011-11-11T16:34:44 and all I do is get food 2011-11-11T16:35:44 *** retybok_ has quit IRC (Ping timeout: 256 seconds) 2011-11-11T16:36:27 I'm at #163 after 12 games, with pathfinding thats cheap enough not to have implemented timeout code yet :-) 2011-11-11T16:36:55 same, my pathfinding is just 1 bfs per turn lol 2011-11-11T16:38:23 *** bergmark has quit IRC (Ping timeout: 248 seconds) 2011-11-11T16:53:23 *** jasox has joined #aichallenge 2011-11-11T16:55:13 *** Anilm3 has quit IRC (Ping timeout: 265 seconds) 2011-11-11T16:55:31 *** Antimony has joined #aichallenge 2011-11-11T16:58:07 *** Zaphus has quit IRC (Ping timeout: 265 seconds) 2011-11-11T16:58:57 *** dwins has quit IRC (Ping timeout: 252 seconds) 2011-11-11T17:00:37 *** ccc_ has quit IRC (Quit: Page closed) 2011-11-11T17:00:52 *** goffrie_ is now known as goffrie 2011-11-11T17:00:59 *** dwins has joined #aichallenge 2011-11-11T17:01:07 no pathfinder, around 1500 2011-11-11T17:01:23 just a build in magnet 2011-11-11T17:04:45 *** Antimony has quit IRC (Ping timeout: 260 seconds) 2011-11-11T17:15:55 just out of curiosity: i play a game. my mu is recalculated. now, do the rankings of all 10k players have to be reassigned, or is there some clever trick to do that on a subset ? 2011-11-11T17:19:17 g0llum: The ranking is just a sorted list of skill 2011-11-11T17:19:35 so that's all of 10k 2011-11-11T17:19:57 the algorithm uses everyone as input and generates the skills for everyone 2011-11-11T17:20:06 Well, it's all of the ones ranked lower than yours 2011-11-11T17:20:15 But that's pretty straightforward to do - just a sort operation 2011-11-11T17:20:36 trueskill 2011-11-11T17:20:44 ah, you can skip the lower ones.. 2011-11-11T17:20:58 ...not really? 2011-11-11T17:21:11 more of a database question 2011-11-11T17:21:41 i think it's just using a completely new ranking 2011-11-11T17:22:50 You can update incrementally if you use a tree structure 2011-11-11T17:23:04 But sorting 10k elements isn't really that hard 2011-11-11T17:23:42 k, seems i'm just scared by numbers here 2011-11-11T17:24:02 Much less hard than simulating 10k^2 games :) 2011-11-11T17:24:22 shure ;) 2011-11-11T17:24:54 * avdg wonders if a bot could win by breaking the algorithm 2011-11-11T17:25:10 it would be fun 2011-11-11T17:25:27 avdg: I don't see how 2011-11-11T17:25:41 me neither 2011-11-11T17:25:43 just wondering 2011-11-11T17:25:51 The inputs are the outcomes of the games, and it's fairly straightforward; the best way to break it would be to consistently win. 2011-11-11T17:28:29 *** Flapadar has joined #aichallenge 2011-11-11T17:32:56 *** bergmark has joined #aichallenge 2011-11-11T17:33:48 *** Anilm3 has joined #aichallenge 2011-11-11T17:35:17 *** conor_f has quit IRC (Quit: leaving) 2011-11-11T17:36:04 How large can the maps be? 2011-11-11T17:36:09 *** tobiassjosten has quit IRC (Ping timeout: 258 seconds) 2011-11-11T17:37:02 *** djr_ has joined #aichallenge 2011-11-11T17:37:32 200x200, but the largest yet are not bigger than 150x150 2011-11-11T17:37:45 *** Flapadar has quit IRC (Ping timeout: 265 seconds) 2011-11-11T17:38:16 rofer, don#t be t on anything, a 1000*1000 game would trigger epilepsy on the hunan watching it, but it's not impossible.. 2011-11-11T17:39:04 g0llum: was worried about that, thanks 2011-11-11T17:39:46 the spec says games can't be larger than 200x200 (just looked up again to be sure) 2011-11-11T17:40:00 http://aichallenge.org/specification.php <- lookup for "map generators" 2011-11-11T17:40:23 the fact, that some organizers are still playing around with mapbuilders should say something, too.. 2011-11-11T17:40:52 breaking the spec would be the last thing organizers would do ;-) 2011-11-11T17:41:17 *** Antimony has joined #aichallenge 2011-11-11T17:41:28 but maybe they could do it for fun 2011-11-11T17:41:41 but supplying a complete set ofmaps 2 days before final submission ? i bet on that ! 2011-11-11T17:42:03 there are more important things to break first 2011-11-11T17:42:10 new maps of course 2011-11-11T17:42:32 *** tobiassjosten has joined #aichallenge 2011-11-11T17:44:34 *** ztfw has quit IRC (Remote host closed the connection) 2011-11-11T17:48:22 *** Antimony has quit IRC (Ping timeout: 255 seconds) 2011-11-11T17:49:11 *** dwins has quit IRC (Quit: Lost terminal) 2011-11-11T17:53:44 http://ants.fluxid.pl/replay.30714 starting to figure out a way to defend :) 2011-11-11T17:55:53 can't help it, took sides with bugnuts there.. 2011-11-11T17:58:24 turn 460 , 1 against 15, still standing 2011-11-11T17:58:54 yeah, he has 64 ants waiting to spawn 2011-11-11T17:59:13 aww, that explains it 2011-11-11T17:59:33 I added code to stop my ants from stepping on a hill with ants spawning out of it, but it's counterproductive in this case b/c he is moving his ants off the hill right away 2011-11-11T17:59:38 so I need to fix that with a heuristic 2011-11-11T18:00:05 explain ;) 2011-11-11T18:00:29 well, if you step onto their hill and they spawn an ant, your ant dies and you don't raze the hill 2011-11-11T18:00:43 but if they immediately move the ant off the hill I think you can still raze it by moving onto it 2011-11-11T18:00:53 i.e. they cooperate with the razing 2011-11-11T18:01:23 he's moving off the hill right away and colliding with my ants, so my current tactic of waiting for the hill to finish spawning ants failed 2011-11-11T18:02:17 *** Antimony has joined #aichallenge 2011-11-11T18:03:04 niether you or him know, that there are 64 more inside that hill 2011-11-11T18:03:41 yeah, but atm I try to surround the hill, and I won't move onto it if an ant spawned there last turn basically 2011-11-11T18:04:11 if you zoom in and step through turn-by-turn you can see he's stepping off the hill and colliding with my ants rather than staying there 2011-11-11T18:05:06 *** djr_ has quit IRC (Read error: Connection reset by peer) 2011-11-11T18:05:48 I'm not sure how to counter that :\ 2011-11-11T18:06:36 *** djr_ has joined #aichallenge 2011-11-11T18:08:19 *** Flapadar has joined #aichallenge 2011-11-11T18:09:24 *** RobotCaleb has quit IRC (Ping timeout: 276 seconds) 2011-11-11T18:11:14 *** treeform has quit IRC (Remote host closed the connection) 2011-11-11T18:11:47 *** treeform has joined #aichallenge 2011-11-11T18:12:59 maybe the dead ants thing could help with this... 2011-11-11T18:13:50 maybe you need a secondary mayor in psycology for that 2011-11-11T18:14:10 or just a lucky hand 2011-11-11T18:14:56 I think you get told about dead ants in a tile 2011-11-11T18:15:09 so I could check and see if there are newly dead enemies in the 4 squares around the hill 2011-11-11T18:15:49 like, front is coming closer ? 2011-11-11T18:16:20 well, I need to know if I should surround the hill and wait for it to drain or move onto it right away 2011-11-11T18:16:31 if they're leaving their newly spawned ants on the hill, surrounding it is best 2011-11-11T18:16:47 if they're doing kamikaze collision attacks, moving onto it right away will work 2011-11-11T18:17:08 and hopefully their tactic doesn't change :P 2011-11-11T18:17:13 *** yahim91 has joined #aichallenge 2011-11-11T18:18:03 ah, I can check if the last ant that spawned on the hill stayed there and died 2011-11-11T18:18:35 Does anyone know what the field isTraced is for on the C++ pack? 2011-11-11T18:19:38 On the Square struct... 2011-11-11T18:19:40 *** Antimony_ has joined #aichallenge 2011-11-11T18:19:51 *** delt0r_ has quit IRC (Ping timeout: 258 seconds) 2011-11-11T18:19:58 Anilm3: https://github.com/aichallenge/aichallenge/blob/epsilon/ants/dist/starter_bots/cpp/Square.h I don't see it 2011-11-11T18:20:14 do you have an old version? 2011-11-11T18:20:16 *** Fandekasp has quit IRC (Ping timeout: 248 seconds) 2011-11-11T18:20:33 Probably 2011-11-11T18:20:36 Thanks anyway 2011-11-11T18:20:37 *** dorisabayon has quit IRC (Ping timeout: 258 seconds) 2011-11-11T18:20:54 it's not in any of the git revisions 2011-11-11T18:21:01 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-11-11T18:21:05 *** Antimony_ is now known as Antimony 2011-11-11T18:22:07 I dont know where it came from 2011-11-11T18:22:49 Maybe I did it when I started... Thanks anyway thestinger 2011-11-11T18:23:49 Anyone know how to tell what is a hill in Lua? There's no type specified for it in the examples. Only land,dead,ant,water and food. 2011-11-11T18:25:36 *** Antimony has quit IRC (Ping timeout: 258 seconds) 2011-11-11T18:26:19 i don't understand how to tell if a tile is wall? 2011-11-11T18:27:24 *** Jak_o_Shadows has joined #aichallenge 2011-11-11T18:27:37 *** lesp has joined #aichallenge 2011-11-11T18:27:52 why does google maps always go to north america by default? 2011-11-11T18:28:03 why don't they go to where they think the user is based off ip location? 2011-11-11T18:28:28 they do after all base search results somewhat off a users location yeah? 2011-11-11T18:28:31 *** foRei has quit IRC (Quit: Bye) 2011-11-11T18:29:51 *** lesp has quit IRC (Client Quit) 2011-11-11T18:32:38 antimatroid: you can set a default location 2011-11-11T18:32:42 *** Fandekasp has joined #aichallenge 2011-11-11T18:32:45 a1k0n: not the point :P 2011-11-11T18:32:50 *** dorisabayon has joined #aichallenge 2011-11-11T18:32:54 that is a good question 2011-11-11T18:32:59 are search results not filtered by ip location regardless of what the user sets? 2011-11-11T18:33:09 *** delt0r_ has joined #aichallenge 2011-11-11T18:33:09 they seem to be, yes 2011-11-11T18:33:20 perhaps when i go back to work i will find out 2011-11-11T18:33:24 i've stolen this from askreddit, but i find it curious 2011-11-11T18:33:33 a1k0n: i would be very interested in the answer :) 2011-11-11T18:33:42 *** choas has quit IRC (Ping timeout: 240 seconds) 2011-11-11T18:33:57 lol 2011-11-11T18:33:59 although i am going to be gone for a week once you get to monday, @later tell me the answer :) 2011-11-11T18:34:10 i have to fight it to get north american results 2011-11-11T18:34:13 lucky bastard 2011-11-11T18:34:21 what? 2011-11-11T18:34:27 i won't be back at work for a few weeks 2011-11-11T18:34:40 ah easy then 2011-11-11T18:37:12 *** Anilm3 has quit IRC (Ping timeout: 265 seconds) 2011-11-11T18:37:39 *** Anilm3 has joined #aichallenge 2011-11-11T18:39:07 *** ajf has quit IRC (Quit: leaving) 2011-11-11T18:40:27 antimatroid or anyone else with interest, any comments on these proposed rule changes? https://github.com/aichallenge/aichallenge/issues/344 2011-11-11T18:41:46 janzert: sounds good to me 2011-11-11T18:42:09 ok, thanks 2011-11-11T18:42:58 janzert: do you have experience with unionfs? 2011-11-11T18:43:10 aaah finally, my A* function now only takes <1ms :D 2011-11-11T18:43:11 @later tell amstan if you approve I'm going to make the proposed rule changes on the site 2011-11-11T18:43:11 janzert: Yes master! 2011-11-11T18:43:15 amstan: some 2011-11-11T18:43:43 janzert: i'm ok with those rules, should prevent some of the questions 2011-11-11T18:43:44 and should have not made that an @later obvioiusly :) 2011-11-11T18:43:57 hopefully :) 2011-11-11T18:44:34 janzert: i'm just looking at unionfs for media organization, i basically want RAID0 without the risk of loosing all the files even if they're on unaffected drives 2011-11-11T18:44:36 janzert: lgtm 2011-11-11T18:44:48 (looks good to me) 2011-11-11T18:45:56 hmm, you might be able to make it work although all writes will go to one specified spot of course 2011-11-11T18:46:03 Zannick: great 2011-11-11T18:46:21 janzert: i think i'll write to the individual nodes, i just want to have an all in one reading experience 2011-11-11T18:47:01 also it seems the only currently stable and maintained version right now is the fusefs one 2011-11-11T18:47:18 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-11-11T18:48:02 *** Harpyon has quit IRC (Quit: Textual IRC Client: http://www.textualapp.com/) 2011-11-11T18:49:02 *** mcstar has left #aichallenge ("WeeChat 0.3.6") 2011-11-11T18:52:42 aichallenge: janzert epsilon * rd7339b5 / website/rules.php : Update and clarify contest rules - http://git.io/nCB05g 2011-11-11T18:53:17 *** sigh has joined #aichallenge 2011-11-11T18:53:44 I made some minor grammar updates from my last post in the issue so you might want review the changes 2011-11-11T18:54:08 *** Harpyon has joined #aichallenge 2011-11-11T18:57:24 *** Anilm3 has quit IRC (Quit: Page closed) 2011-11-11T18:59:34 janzert: looks alright 2011-11-11T19:02:53 *** lavalamp has quit IRC (Quit: Page closed) 2011-11-11T19:09:16 *** Pythonoob has joined #aichallenge 2011-11-11T19:09:53 Hi all 2011-11-11T19:10:27 *** Pythonoob has quit IRC (Client Quit) 2011-11-11T19:18:39 *** Jak_o_Shadows has quit IRC (Remote host closed the connection) 2011-11-11T19:21:39 *** jasox has quit IRC (Quit: Leaving) 2011-11-11T19:22:14 *** Fandekasp has quit IRC (Ping timeout: 260 seconds) 2011-11-11T19:22:20 *** dorisabayon has quit IRC (Ping timeout: 258 seconds) 2011-11-11T19:23:02 *** smiley1983 has quit IRC (Ping timeout: 260 seconds) 2011-11-11T19:24:09 *** smiley1993 has quit IRC (Ping timeout: 276 seconds) 2011-11-11T19:25:10 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-11-11T19:28:25 *** Zaphus has joined #aichallenge 2011-11-11T19:29:36 *** smiley1983 has joined #aichallenge 2011-11-11T19:30:29 *** smiley1993 has joined #aichallenge 2011-11-11T19:30:38 *** Antimony has joined #aichallenge 2011-11-11T19:31:31 *** mleise has joined #aichallenge 2011-11-11T19:35:40 *** grwip has quit IRC (Quit: Leaving) 2011-11-11T19:35:51 *** Antimony_ has joined #aichallenge 2011-11-11T19:36:14 *** Harpyon has quit IRC (Quit: Computer has gone to sleep.) 2011-11-11T19:37:35 *** Antimony has quit IRC (Ping timeout: 258 seconds) 2011-11-11T19:37:46 *** Antimony_ is now known as Antimony 2011-11-11T19:37:54 *** ikaros has joined #aichallenge 2011-11-11T19:40:19 *** Antimony has quit IRC (Client Quit) 2011-11-11T19:48:44 *** t2027 has quit IRC (Ping timeout: 265 seconds) 2011-11-11T19:49:15 *** Conorach has quit IRC (Ping timeout: 260 seconds) 2011-11-11T19:53:55 *** dserrano3 has joined #aichallenge 2011-11-11T19:55:07 *** dserrano3 has quit IRC (Client Quit) 2011-11-11T19:55:38 *** Zaphus has quit IRC (Quit: Page closed) 2011-11-11T20:00:20 *** ProperChaos has quit IRC (Ping timeout: 265 seconds) 2011-11-11T20:05:36 *** lvbreda has joined #aichallenge 2011-11-11T20:06:12 *** amstan has quit IRC (Read error: Connection reset by peer) 2011-11-11T20:06:26 *** amstan has joined #aichallenge 2011-11-11T20:06:26 *** ChanServ sets mode: +o amstan 2011-11-11T20:07:15 Does anybody elsehave problems with the map maze_02p_02 only when Itry this map I get a python error ' while the bot is written in java ' and itis only sometimes. 2011-11-11T20:08:20 *** Transformer has joined #aichallenge 2011-11-11T20:09:36 *** Transformer has quit IRC (Excess Flood) 2011-11-11T20:14:20 *** amstan has quit IRC (Ping timeout: 260 seconds) 2011-11-11T20:14:39 janzert: I added a comment to the contest rules issue you mentoned 2011-11-11T20:16:17 *** regularguy has quit IRC (Ping timeout: 265 seconds) 2011-11-11T20:16:43 yeah, eventually moving to cpu time would be nice although too late for this contest and we need someone to write the code to switch over to it. Basically we're still on wall time because no one has written the cpu time tracking 2011-11-11T20:16:59 Ok, what would be the deadline for the next contest? 2011-11-11T20:17:30 so reviewing my recent losses I just realized I may have lost this replay on the very first turn: http://aichallenge.org/visualizer.php?game=83511&user=3473 2011-11-11T20:17:33 there are no plans regarding the next one yet, but I would say a minimum of 3 months after this one 2011-11-11T20:17:51 more probably a 6 months to a year 2011-11-11T20:17:59 there's food right by the hill and I step away 2011-11-11T20:18:11 oops :( 2011-11-11T20:20:14 at least it's easy to fix! 2011-11-11T20:20:54 yeah, and should help any other situation where food spawns right next to an ant 2011-11-11T20:21:20 I probably have a specific exception for that which was just wrong thinking 2011-11-11T20:22:23 oh ho, I know what it is 2011-11-11T20:22:31 my old nemesis treating my own hills as impassible to avoid stepping on them 2011-11-11T20:26:46 *** TheLinker has joined #aichallenge 2011-11-11T20:27:15 *** datachomper has joined #aichallenge 2011-11-11T20:40:05 *** datachomper has joined #aichallenge 2011-11-11T20:44:46 *** datachomper has quit IRC (Client Quit) 2011-11-11T20:54:20 *** sigh has quit IRC (Remote host closed the connection) 2011-11-11T20:58:18 *** NoxiaZ^ has quit IRC (Ping timeout: 245 seconds) 2011-11-11T20:58:18 *** treeform_ has joined #aichallenge 2011-11-11T21:02:19 *** treeform has quit IRC (Ping timeout: 255 seconds) 2011-11-11T21:04:04 *** Relio has joined #aichallenge 2011-11-11T21:05:04 *** JorgeB has quit IRC (Ping timeout: 258 seconds) 2011-11-11T21:09:17 *** treeform_ has quit IRC (Remote host closed the connection) 2011-11-11T21:12:13 *** sigh has joined #aichallenge 2011-11-11T21:14:00 *** lvbreda has quit IRC (Quit: Page closed) 2011-11-11T21:17:04 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-11-11T21:21:06 *** ccc has joined #aichallenge 2011-11-11T21:22:34 *** liang has joined #aichallenge 2011-11-11T21:28:06 *** ccc has quit IRC (Quit: Computer has gone to sleep.) 2011-11-11T21:33:11 *** treeform has joined #aichallenge 2011-11-11T21:35:11 *** treeform has quit IRC (Remote host closed the connection) 2011-11-11T21:39:13 Anyone know why the go starter package log messages don't appear in the output of the playgame script? 2011-11-11T21:39:24 *** yahim91 has quit IRC (Quit: Page closed) 2011-11-11T21:39:25 Python log messages do, and I've tried explicitly setting the output writer as stderr, too. 2011-11-11T21:39:37 you have to pass -e to playgame.py 2011-11-11T21:39:43 otherwise it throws away stderr 2011-11-11T21:40:11 *** Dabrion has quit IRC (Quit: Page closed) 2011-11-11T21:46:55 ah, thanks 2011-11-11T21:51:58 This is perplexing: it's crashing because I'm trying to move an ant onto a square occupied by another ant 2011-11-11T21:52:09 but that should never happen - I'm checking if the square is safe first 2011-11-11T21:52:43 nickjohnson: what about moving two ants to the same square? you sure that isn't happening? 2011-11-11T21:53:12 Extrarius: The Go starter kit has a 'SafeDestination' function that checks for ants; I'm skipping any squares for which it returns False 2011-11-11T21:53:31 On top of that, I've seen frames where it let me step into an occupied square without warning or complaint 2011-11-11T22:10:13 *** nikis has quit IRC (Ping timeout: 240 seconds) 2011-11-11T22:11:03 Oops, I'm an idiot 2011-11-11T22:11:05 *** nikis has joined #aichallenge 2011-11-11T22:13:34 *** lknix has quit IRC (Quit: WeeChat 0.3.6) 2011-11-11T22:14:52 *** TheLinker has quit IRC (Remote host closed the connection) 2011-11-11T22:23:17 *** liang has quit IRC (Quit: Page closed) 2011-11-11T22:24:45 *** thestinger has quit IRC (Quit: WeeChat 0.3.6) 2011-11-11T22:37:25 *** Flapadar has quit IRC (Ping timeout: 265 seconds) 2011-11-11T22:41:31 *** thestinger has joined #aichallenge 2011-11-11T22:44:38 *** janzert has quit IRC (Read error: Connection reset by peer) 2011-11-11T22:47:25 *** smiley1983 has quit IRC (Ping timeout: 260 seconds) 2011-11-11T22:47:32 *** janzert has joined #aichallenge 2011-11-11T22:48:35 *** ninjax has joined #aichallenge 2011-11-11T22:49:07 *** smiley1983 has joined #aichallenge 2011-11-11T22:50:55 How many people are using potential fields rather than path finding? 2011-11-11T22:53:15 *** pedrosorio has quit IRC (Quit: Page closed) 2011-11-11T22:54:38 *** McLeopold has joined #aichallenge 2011-11-11T22:59:10 *** Hexren1 has quit IRC (Quit: Leaving.) 2011-11-11T22:59:57 *** McLeopold has quit IRC (Quit: Page closed) 2011-11-11T23:01:22 *** ninjax has quit IRC (Remote host closed the connection) 2011-11-11T23:02:10 *** ninjax has joined #aichallenge 2011-11-11T23:03:01 *** datachomper has joined #aichallenge 2011-11-11T23:05:06 *** ninjax has quit IRC (Remote host closed the connection) 2011-11-11T23:06:14 * nickjohnson is 2011-11-11T23:10:32 I'm using influence map as input for pathfinding 2011-11-11T23:11:35 'influence map'? 2011-11-11T23:23:25 *** fusiongenex has joined #aichallenge 2011-11-11T23:27:50 *** datachomper has quit IRC (Quit: Leaving.) 2011-11-11T23:31:42 *** fusiongenex has quit IRC (Quit: Page closed) 2011-11-11T23:45:50 *** Slin has joined #aichallenge 2011-11-11T23:46:42 Hello. I was wondering if anyone knew where/if I can find the old page with the final rankings and profiles for the planet wars contest? 2011-11-11T23:52:53 *** Slin has quit IRC (Quit: Page closed) 2011-11-11T23:55:07 *** antimatroid has quit IRC (Quit: Leaving.) 2011-11-11T23:55:15 *** antimatroid has joined #aichallenge 2011-11-11T23:55:52 *** b0rder has joined #aichallenge 2011-11-11T23:57:14 *** Angela has joined #aichallenge 2011-11-11T23:57:40 how to let the ants explores map automatically? 2011-11-11T23:57:40 *** Angela is now known as Guest19284 2011-11-11T23:59:04 could you give me some ideas for explore map?