2010-11-04T00:01:10 ... 2010-11-04T00:02:56 *** antimatroid1 has joined #aichallenge 2010-11-04T00:04:32 *** antimatroid has quit IRC (Ping timeout: 240 seconds) 2010-11-04T00:08:29 *** virdo has joined #aichallenge 2010-11-04T00:14:53 Top 10 players: bocsimacko(4231), bix0r4ever(4106), jimrogerz(4068), dmj111(4065), george(4052), shangas(4013), GreenTea(4013), Raschi(3977), Hazard(3957), Accoun(3939) 2010-11-04T00:25:02 *** janzert has quit IRC (Read error: Connection reset by peer) 2010-11-04T00:26:34 *** janzert has joined #aichallenge 2010-11-04T00:29:56 lol been checkign all my state calcs etc thinking it wasn't updating right 2010-11-04T00:30:04 i had the eval func before the while loop lol 2010-11-04T00:30:28 i can't get my new moves to run without crashing 2010-11-04T00:31:44 yea sorry to all the people i've crashed on today 2010-11-04T00:31:46 been a rough one 2010-11-04T00:32:10 i did just put in c compliance today expected a few new bugs 2010-11-04T00:32:14 but there were some old ones too hah 2010-11-04T00:32:40 i'm not playing on tcp while it crashes 2010-11-04T00:33:31 haha well i do keep thinking i have it fixed 2010-11-04T00:33:41 not like i just go on there knowing i'm crashing ;) i try to keep it local for that 2010-11-04T00:40:25 that doesn't even make sense 2010-11-04T00:40:42 i have the contents of a function entirely commented out, if i call it it crashes, if i don't it doesn't 2010-11-04T00:41:52 sounds like someone is playing around in memory that he shouldn't 2010-11-04T00:43:27 hmmm 2010-11-04T00:43:31 c++ is good at letting me do that 2010-11-04T00:43:35 finding those problems are a bitch 2010-11-04T00:43:46 have you got any tools like valgrind? 2010-11-04T00:44:03 nope :P 2010-11-04T00:44:33 i'll check stuff iterating through objects 2010-11-04T00:44:46 *** hellman has joined #aichallenge 2010-11-04T00:46:17 seriously going c-compliant was a good move on my part 2010-11-04T00:46:31 you don't have to go that far obv but man i've found a lot of shit 2010-11-04T00:46:50 demerzel__: what did you find? 2010-11-04T00:47:12 well for one (that isn't necessarily c-compliance) - i access struct arrays directly which helps a ton 2010-11-04T00:47:24 and forcing pointers over ref has saved me a lot of scoping trouble 2010-11-04T00:47:42 doh, it would help if i commented out implementMove rather than implementNewMove aha 2010-11-04T00:47:48 no member methods and extern globals means i know where everything is at all times haha 2010-11-04T00:48:03 antimatroid1: haha 2010-11-04T00:48:09 yea probably lol 2010-11-04T00:48:18 *** tapwater has quit IRC (Quit: tapwater) 2010-11-04T00:48:38 demerzel__: yeah, removing globals always helps... not sure how member methods are harmful though 2010-11-04T00:49:01 well on structs they aren't 2010-11-04T00:49:02 *** Mathnerd314 has quit IRC (Quit: ChatZilla 0.9.86-rdmsoft [XULRunner 1.9.2.11/20101012113537]) 2010-11-04T00:49:05 i take that back 2010-11-04T00:49:17 idk there are a lot of scoping problems with them even static methods do funky shit 2010-11-04T00:49:26 and if you're gonna go static it doesn't hurt to make them global 2010-11-04T00:49:51 i spent more time tracking down vars than actually fixing calcs 2010-11-04T00:49:55 now i don't have that problem :) 2010-11-04T00:50:08 hmm... ok 2010-11-04T00:50:35 *** virdo has quit IRC (Ping timeout: 252 seconds) 2010-11-04T00:56:46 j3camero, janzert: ping 2010-11-04T00:56:49 *** computerwiz_222 has joined #aichallenge 2010-11-04T00:56:54 pong 2010-11-04T00:57:12 janzert: how would you recomend amazon ec2? 2010-11-04T00:57:32 in what way? 2010-11-04T00:57:48 It's great for what it is 2010-11-04T00:58:17 I don't think I'd use it in place of say a dedicated server for a web site or such though 2010-11-04T00:58:27 well, i'm thinking of getting a webapp, it needs lots of storage, cpu usage and bandwidth 2010-11-04T00:59:24 the storage through s3 is probably good for quite a few situations 2010-11-04T01:00:00 for cpu usage if it is sporadic ec2 may be great 2010-11-04T01:00:30 for bandwidth if you can use the CDN it looks pretty good as well 2010-11-04T01:00:38 janzert: i'm thinking of starting with this for now: http://aws.amazon.com/free/ 2010-11-04T01:00:45 yeah, i'll probably have to use CDN 2010-11-04T01:01:20 that free package does look like a nice way to try it out 2010-11-04T01:01:53 notice that the micro instances are pretty limited and only use EBS storage (which I have no experience with) 2010-11-04T01:02:02 *** JamesGecko__ has joined #aichallenge 2010-11-04T01:02:39 so i won't be getting the CDN with this one then? 2010-11-04T01:03:04 you can use it, it just won't be free 2010-11-04T01:03:32 the free "package" isn't a real package per se 2010-11-04T01:03:46 it's is just some initial free use you are allowed 2010-11-04T01:03:57 anything beyond those is just charged at the normal rates 2010-11-04T01:05:06 *** JamesGecko_ has quit IRC (Ping timeout: 240 seconds) 2010-11-04T01:05:10 also, how would one go about hosting the http frontend on it, do i just point my domain to their dns? 2010-11-04T01:06:39 You would have to set up some sort of dns hosting either yourself or use one of the free ones around 2010-11-04T01:06:48 i can do that then 2010-11-04T01:06:51 and point it at your instance 2010-11-04T01:06:53 but don't the ips change? 2010-11-04T01:07:44 I've only really used spot request instances that don't hang around long, but... 2010-11-04T01:08:09 it seems that the IP is permanently assigned to the instance as long as you keep it 2010-11-04T01:08:21 cool 2010-11-04T01:08:27 well, i'm sold 2010-11-04T01:09:35 it looks like you can also get an "elastic ip" which is a static ip assigned to your account that you can use as you want 2010-11-04T01:09:51 janzert: can i get multiple users per host? 2010-11-04T01:10:20 you have root, so you can set it up just like you would a dedicated server basically 2010-11-04T01:10:30 awesome 2010-11-04T01:11:09 *** TheMacca has joined #aichallenge 2010-11-04T01:12:35 *** tapwater has joined #aichallenge 2010-11-04T01:12:36 of course one thing to be aware of is that you probably get more support from a traditional provider if you need help running your server 2010-11-04T01:12:54 nah, i know my way around running servers 2010-11-04T01:14:54 Top 10 players: bocsimacko(4229), jimrogerz(4132), bix0r4ever(4108), george(4060), dmj111(4050), shangas(4007), GreenTea(4005), Accoun(3977), Raschi(3954), Hazard(3951) 2010-11-04T01:17:50 nice thanks for the free tip 2010-11-04T01:21:04 *** wh1teside has joined #aichallenge 2010-11-04T01:24:13 *** wh1teside_ has quit IRC (Ping timeout: 255 seconds) 2010-11-04T01:26:22 stupid bug was stupid 2010-11-04T01:26:33 i was passing the wrong variable into a constructor for fleets 2010-11-04T01:34:22 *** computerwiz_222 has quit IRC (Quit: computerwiz_222) 2010-11-04T01:34:30 holy shit i think i finally fixed it 2010-11-04T01:34:35 was the distance order sorting haha 2010-11-04T01:34:44 had dupes so it would just loop forever trying to find the closest source 2010-11-04T01:42:42 *** nullkuhl has quit IRC (Read error: Connection reset by peer) 2010-11-04T01:44:17 *** nullkuhl has joined #aichallenge 2010-11-04T01:45:53 haha... today the channel is just feed of all the bugs people have fixed :P 2010-11-04T01:50:53 lol 2010-11-04T01:50:57 i am still trying to track one down 2010-11-04T01:51:07 some combination of indices and distances breaks my distance sort :( 2010-11-04T01:52:13 just us two :P 2010-11-04T01:52:32 my bugs are gone, now i need to make it smart 2010-11-04T01:52:39 i guarantee my bugs were more numerous 2010-11-04T01:52:58 stitched together a c++ bot from rusty knowledge and then transfered it to c lol 2010-11-04T01:54:00 haha: "antimatroid1: my bugs are gone" 2010-11-04T01:54:29 i say that and it just crashed :P 2010-11-04T01:54:34 haha 2010-11-04T01:54:36 :P 2010-11-04T01:54:56 i was referring to the bugs that make it crash, semantic bugs are still plentiful i think 2010-11-04T01:55:26 heheh, new updates on my terrain generator: http://cubicmatter.net/forums/viewtopic.php?f=10&t=54#p454 2010-11-04T01:57:11 what is cube2 2010-11-04T01:57:55 demerzel__: this: http://sauerbraten.org/ 2010-11-04T01:58:53 does it have physics 2010-11-04T01:59:05 we have now pass the 4000 active submissions mark 2010-11-04T01:59:08 oh it's an fps 2010-11-04T01:59:21 @yay 2010-11-04T01:59:21 Double rainbow all the way! 2010-11-04T01:59:58 what counts as an inactive submission? 2010-11-04T02:00:00 janzert: but how many don't suck 2010-11-04T02:00:09 1 2010-11-04T02:00:22 yeah, pretty much :P 2010-11-04T02:00:24 you mean 4000 + x -1? 2010-11-04T02:00:31 oh wait, nevermind 2010-11-04T02:00:35 i'm having a bad day :P 2010-11-04T02:00:55 stop throwing equations at me, it hurts 2010-11-04T02:00:57 *** wh1teside has quit IRC (Ping timeout: 245 seconds) 2010-11-04T02:01:09 a grade 5 student could handle that :P 2010-11-04T02:01:26 I'm not a grade 5 student? 2010-11-04T02:01:40 *** narnach has joined #aichallenge 2010-11-04T02:01:48 that was a statement, not a question 2010-11-04T02:02:14 despite initial appearances 2010-11-04T02:03:30 *** boegel has joined #aichallenge 2010-11-04T02:04:19 janzert: has the elo offset been changed? top elo are well over 4000 but lowest elo are not negative yet 2010-11-04T02:05:56 no, the offset is still the same 2010-11-04T02:06:13 unless someone else snuck in a change and didn't commit it back to svn :} 2010-11-04T02:06:20 haha 2010-11-04T02:06:33 I keep waiting for negative elo :P 2010-11-04T02:06:47 yeah, I was a bit surprised that it's still positive as well 2010-11-04T02:07:18 the offset is 2000 to the best of your knowledge, yeah? 2010-11-04T02:07:24 right 2010-11-04T02:07:48 *** JamesGecko_ has joined #aichallenge 2010-11-04T02:08:04 when we do hit negative I'm thinking of anchoring the bottom instead of having an added offset 2010-11-04T02:09:17 *** boegel has quit IRC (Quit: *poof!*) 2010-11-04T02:09:31 some poor sod is going to look at it and think they're worse than nothing :P 2010-11-04T02:09:47 hehe 2010-11-04T02:10:02 that's why I'd anchor it at either 0 or maybe 100 2010-11-04T02:10:12 or some such 2010-11-04T02:10:24 or you can just do some ridiculously high offset 2010-11-04T02:10:42 make it like 7000 2010-11-04T02:10:47 that way the top bots can be over 9000 :P 2010-11-04T02:11:04 *** JamesGecko__ has quit IRC (Ping timeout: 252 seconds) 2010-11-04T02:11:14 :) at least anchoring it has some rationalization besides just making people happy 2010-11-04T02:11:17 why not 1 billion? :P 2010-11-04T02:11:52 janzert: what's the rationalisation? 2010-11-04T02:12:47 all the bottom bots are instant loss and it makes some sense that they should have a "defined" elo from which everything else builds up from 2010-11-04T02:13:16 zero seems like the best but might be avoided for feelings 2010-11-04T02:13:16 that actually makes sense 2010-11-04T02:14:01 0 is fine i think... I think people know that their bot is not that great if it loses every game on the first turn 2010-11-04T02:14:02 gotdam found another one that was supposed to prevent crashes actually made them :( 2010-11-04T02:14:13 my bot could have a 7000 score and it'd still be shit 2010-11-04T02:14:54 Top 10 players: bocsimacko(4236), jimrogerz(4135), bix0r4ever(4104), george(4073), dmj111(4057), shangas(4017), GreenTea(4011), Accoun(3976), Raschi(3960), Hazard(3959) 2010-11-04T02:17:29 woop made it through a game w/o crashing lol 2010-11-04T02:18:23 demerzel__: what exactly is causing all these crashes? 2010-11-04T02:18:32 tons of shit 2010-11-04T02:18:39 i mean they're not crashes they're invalids 2010-11-04T02:18:54 there were some loop problems with the order queue tho lol 2010-11-04T02:19:09 ah ok 2010-11-04T02:19:20 i just recoded the entire thing that's basically why all these problems 2010-11-04T02:19:36 but i think it's mostly under control now i think maybe 1 or 2 lingerers but all the major ones are gone finally 2010-11-04T02:19:45 cool :) 2010-11-04T02:19:54 yea sorry to just bitch into chat 2010-11-04T02:20:01 i had no idea i'd be debugging for a week after writing code for 2 days haha 2010-11-04T02:20:08 :P 2010-11-04T02:20:33 that's just the beginning, the debugging never ends for this 2010-11-04T02:21:09 haha no this has been some intense debugging 2010-11-04T02:21:16 i sure hope not anyway 2010-11-04T02:21:24 finally seems stable ;) 2010-11-04T02:21:37 can't wait to debug tree nodes lol 2010-11-04T02:21:55 the loop problems have shown me i have some impressive calc power tho 2010-11-04T02:22:04 calc power? 2010-11-04T02:22:06 since essentially it was hitting the same node over and over (i only allow 1 level) 2010-11-04T02:22:16 but to hit that node it had to recalc everything 2010-11-04T02:22:23 it spit out like 1.8million records to log in <10 seconds 2010-11-04T02:22:57 with a shitty io logger haha 2010-11-04T02:23:06 *** davidjliu has quit IRC (Quit: Page closed) 2010-11-04T02:23:45 i hope that means i can brute force a few levels 2010-11-04T02:23:49 like i planned 2010-11-04T02:23:51 we'll see :) 2010-11-04T02:24:23 errr.... what do you count as "brute force" 2010-11-04T02:25:01 not using a/b 2010-11-04T02:25:27 yeah, but brute force implies you are doing some sort of exhaustive search 2010-11-04T02:25:42 what exactly are you exhaustively searching? 2010-11-04T02:25:51 move hierarchies 2010-11-04T02:26:13 or trees if you prefer 2010-11-04T02:26:36 i am still defining what a move will be tho 2010-11-04T02:26:37 sure, but I doubt you are searching all the moves.... 2010-11-04T02:26:40 yeah, fair enough 2010-11-04T02:26:47 well all the moves would be 1-X ships to every other planet 2010-11-04T02:27:03 i did try that on my first bot it did ok but was wasted most of the time 2010-11-04T02:27:12 can at least lowerbound to ships+1 or defense 2010-11-04T02:27:44 even lower bounding gives a ridiculous number 2010-11-04T02:27:45 mostly i am tryign to figure out if i want to consider trees as sequential or immediate 2010-11-04T02:28:21 i managed 1-x in C# - granted only 1 ply but that was close to 2000 * 23! 2010-11-04T02:29:04 wow, I'm actually suprised you got 1 ply 2010-11-04T02:29:29 i optimized the shit out of it but it never timed out 2010-11-04T02:29:35 *** JamesGecko_ has quit IRC (Ping timeout: 252 seconds) 2010-11-04T02:29:42 i dropped that model tho cuz it was my entire turn practically 2010-11-04T02:29:58 not much gain over lowerbounding - if any 2010-11-04T02:30:03 although it auto-defended etc which was nice 2010-11-04T02:31:25 i wonder what mega1 uses for dist 2010-11-04T02:31:33 auto heuristic calc like everybody else or not 2010-11-04T02:31:41 trying to figure out a way to make it part of evals 2010-11-04T02:31:58 anyway, I'm out 2010-11-04T02:32:03 cya 2010-11-04T02:32:04 yea me too :) gn 2010-11-04T02:32:06 *** sigh has quit IRC () 2010-11-04T02:32:20 *** demerzel__ has quit IRC (Quit: Page closed) 2010-11-04T02:36:42 *** rebelxt has joined #aichallenge 2010-11-04T02:37:43 *** mega1 has quit IRC (Ping timeout: 272 seconds) 2010-11-04T02:42:25 *** Utkarsh has quit IRC (Ping timeout: 276 seconds) 2010-11-04T02:42:27 *** Azrathud has quit IRC (Ping timeout: 240 seconds) 2010-11-04T02:47:04 *** Utkarsh has joined #aichallenge 2010-11-04T02:49:02 *** TheMacca has quit IRC (Quit: Page closed) 2010-11-04T03:03:05 *** hightime has joined #aichallenge 2010-11-04T03:13:26 *** Utkarsh has quit IRC (Read error: Connection reset by peer) 2010-11-04T03:14:56 Top 10 players: bocsimacko(4259), jimrogerz(4117), bix0r4ever(4076), george(4053), dmj111(4041), shangas(4000), GreenTea(3991), Accoun(3959), Raschi(3945), Hazard(3938) 2010-11-04T03:20:17 *** Utkarsh has joined #aichallenge 2010-11-04T03:47:46 *** sigh has joined #aichallenge 2010-11-04T03:57:40 *** Appleman1234 has quit IRC (Read error: Connection reset by peer) 2010-11-04T04:02:34 *** Appleman1234 has joined #aichallenge 2010-11-04T04:06:33 lol... my broken indentation made it all the way to the official visualiser.js 2010-11-04T04:13:04 demerzel testingBot is yours? 2010-11-04T04:14:55 Top 10 players: bocsimacko(4298), jimrogerz(4124), bix0r4ever(4089), george(4079), dmj111(4069), shangas(4017), GreenTea(4011), Accoun(3984), rsergio(3970), Hazard(3965) 2010-11-04T04:15:44 *** LEXASOFT has joined #aichallenge 2010-11-04T04:21:56 *** wh1teside has joined #aichallenge 2010-11-04T04:24:28 morning all 2010-11-04T04:25:32 morning 2010-11-04T04:25:42 well its more like 1:30am for me 2010-11-04T04:28:00 lol 2010-11-04T04:28:02 oef 2010-11-04T04:28:09 9:25 here :P 2010-11-04T04:28:24 bhasker: it seems our bots have equal strength atm 2010-11-04T04:28:29 cool 2010-11-04T04:28:31 pretty much yea 2010-11-04T04:28:41 though i am still tinkering with it 2010-11-04T04:28:44 but eventually i will kick ur ass ofc ^^ 2010-11-04T04:28:49 so am i son :P 2010-11-04T04:28:50 hehe you just might 2010-11-04T04:29:01 my bot isn't geting much smarter anymore 2010-11-04T04:29:14 example http://72.44.46.68/canvas?game_id=849487 2010-11-04T04:29:32 ouch 2010-11-04T04:29:45 it seems you win more often of me, but lose more of the others too 2010-11-04T04:30:02 which is really funny 2010-11-04T04:30:08 yea:-\ 2010-11-04T04:30:13 i make some very elementary mistakes 2010-11-04T04:30:16 with staging 2010-11-04T04:30:20 which screws me oer 2010-11-04T04:30:21 over* 2010-11-04T04:30:28 either i overextend on turn 1 2010-11-04T04:30:32 or stage to the wrong planet 2010-11-04T04:30:41 and the moment either of them happens 2010-11-04T04:30:43 i get wtfpwned 2010-11-04T04:32:20 yeah 2010-11-04T04:32:28 staging is very hard to do well 2010-11-04T04:32:43 i'm taking a very naive KISS approach atm 2010-11-04T04:32:47 which is 2010-11-04T04:32:50 which I might wanna look into 2010-11-04T04:33:14 compute the frontline, send ships to closest frontline planet through hubs 2010-11-04T04:33:23 though telling you this stuff is prolly not wise ^^ 2010-11-04T04:33:51 http://72.44.46.68/canvas?game_id=849664 2010-11-04T04:33:53 hah 2010-11-04T04:33:57 same bug 2010-11-04T04:34:03 i really need to factor in ships in future 2010-11-04T04:34:10 i could have sniped the middle 2010-11-04T04:34:14 if i didnt stage 2010-11-04T04:34:26 since i was sending ships i never had enough ships to snipe 2010-11-04T04:34:41 oh man indd 2010-11-04T04:35:33 currently 46 @ official 2010-11-04T04:35:37 been playing all night 2010-11-04T04:35:44 1 loss so far ^^ 2010-11-04T04:35:50 and one in the beginning -.- 2010-11-04T04:35:55 http://ai-contest.com/profile.php?user_id=10154 2010-11-04T04:36:10 I have high hopes :P 2010-11-04T04:36:32 nice 2010-11-04T04:36:58 well it'll also represent ur bot :D 2010-11-04T04:37:14 bhasker: where was your current official ranked at tcp 2010-11-04T04:37:27 so i can get a sort of feeling where this will end up 2010-11-04T04:37:36 #31 odd 2010-11-04T04:37:44 32 2010-11-04T04:37:50 you should probably go higher 2010-11-04T04:37:52 ehh srry wrong question, what elo 2010-11-04T04:38:03 3804 2010-11-04T04:38:03 :) 2010-11-04T04:38:10 no i mean on tcp ^_^ 2010-11-04T04:38:15 oh 2010-11-04T04:38:30 like 5 points above yours 2010-11-04T04:38:43 but my tcp version is different from my official 2010-11-04T04:38:45 so not sure 2010-11-04T04:39:05 yeah i mean where the current official ranked at tcp if it were playing. 2010-11-04T04:39:29 if let say that was around 300 2010-11-04T04:39:41 then i can safely say that i will go quite some points higher 2010-11-04T04:40:03 no idea where it would be 2010-11-04T04:40:12 i don't even remember the revision number of the submission bot lol 2010-11-04T04:40:15 :-\ 2010-11-04T04:40:24 ohw 2010-11-04T04:40:28 use git ^_^ 2010-11-04T04:40:34 http://72.44.46.68/canvas?game_id=849725 you rlly need to fix this 2010-11-04T04:41:28 yea lol 2010-11-04T04:41:30 :( 2010-11-04T04:41:37 stupid thing keeps going for enemy first 2010-11-04T04:41:38 over neutrals 2010-11-04T04:41:50 need to adjust my scoring to factor in support from neutrals 2010-11-04T04:42:04 that way it won't think that it can acquire the enemy 2010-11-04T04:42:06 planet 2010-11-04T04:42:43 i actually created a very sweet/accurate decision system on when to snipe yesterday 2010-11-04T04:42:59 that's one of the reasons i went 80 points up in elo over 16.0 2010-11-04T04:48:21 *** Florian95 has joined #aichallenge 2010-11-04T04:59:21 *** aerique has joined #aichallenge 2010-11-04T05:01:47 *** Appleman1234 has quit IRC (Ping timeout: 265 seconds) 2010-11-04T05:07:11 *** bhasker_ has joined #aichallenge 2010-11-04T05:07:12 *** bhasker has quit IRC (Read error: Connection reset by peer) 2010-11-04T05:09:43 *** Cyndre has quit IRC (Ping timeout: 272 seconds) 2010-11-04T05:14:56 Top 10 players: bocsimacko(4293), jimrogerz(4119), george(4092), bix0r4ever(4084), dmj111(4063), GreenTea(4023), shangas(4018), Accoun(3975), Raschi(3971), rsergio(3955) 2010-11-04T05:16:35 *** espes has quit IRC (Remote host closed the connection) 2010-11-04T05:18:07 *** zerd has quit IRC (Ping timeout: 240 seconds) 2010-11-04T05:26:06 *** Naktibalda has joined #aichallenge 2010-11-04T05:29:21 *** jaspervdj has joined #aichallenge 2010-11-04T05:32:54 7 times "error receiving submission file", what did you do? ^^ 2010-11-04T05:33:13 who me? 2010-11-04T05:33:38 Is anyone else around, except you and bhasker? 2010-11-04T05:33:45 dunno :P 2010-11-04T05:33:58 but yeah, didn't do anything I don't know why that happened 2010-11-04T05:34:15 it was at the time they had huge problems with new submissions 2010-11-04T05:34:40 Frontier: just wait :) 2010-11-04T05:34:54 Yeah, I have a feeling some of these messages just appear, like the bugs when your bot was matched against itself or bots crashed randomly. 2010-11-04T05:38:10 *** Appleman1234 has joined #aichallenge 2010-11-04T05:48:59 does anyone know how galcon does neutral resolution? 2010-11-04T06:07:11 *** Apophis has joined #aichallenge 2010-11-04T06:07:14 *** Arthur has joined #aichallenge 2010-11-04T06:09:11 *** Arthur has quit IRC (Client Quit) 2010-11-04T06:10:20 *** nann has quit IRC (Ping timeout: 252 seconds) 2010-11-04T06:14:35 *** Aaaarg has joined #aichallenge 2010-11-04T06:14:57 Top 10 players: bocsimacko(4293), jimrogerz(4093), george(4079), bix0r4ever(4078), dmj111(4051), GreenTea(4011), shangas(3999), Accoun(3962), Raschi(3960), rsergio(3944) 2010-11-04T06:20:57 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716]) 2010-11-04T06:34:41 *** yasith has quit IRC (Ping timeout: 245 seconds) 2010-11-04T06:38:24 *** mega1 has joined #aichallenge 2010-11-04T06:46:29 *** yasith has joined #aichallenge 2010-11-04T06:47:44 *** Aaaarg has quit IRC (Remote host closed the connection) 2010-11-04T06:55:16 http://72.44.46.68/canvas?game_id=850931 2010-11-04T06:55:19 interesting game 2010-11-04T06:56:04 *** Cold-Phoenix has joined #aichallenge 2010-11-04T07:06:45 *** wh1teside_ has joined #aichallenge 2010-11-04T07:06:45 *** wh1teside has quit IRC (Read error: Connection reset by peer) 2010-11-04T07:11:55 wow, that game was actually entertaining :) 2010-11-04T07:13:28 *** zerd has joined #aichallenge 2010-11-04T07:14:58 Top 10 players: bocsimacko(4298), bix0r4ever(4111), jimrogerz(4109), george(4092), dmj111(4062), shangas(4026), GreenTea(4022), Accoun(3975), felixcoto(3957), Hazard(3955) 2010-11-04T07:19:58 there was another one just after that was good too, hang on 2010-11-04T07:20:09 http://72.44.46.68/canvas?game_id=851016 2010-11-04T07:21:40 the bot isn't doing very well, but it's a first test of a whole new way of picking moves so 2010-11-04T07:27:29 *** wh1teside has joined #aichallenge 2010-11-04T07:30:38 *** wh1teside_ has quit IRC (Ping timeout: 252 seconds) 2010-11-04T07:31:40 *** genericbob has quit IRC (Ping timeout: 276 seconds) 2010-11-04T07:32:53 lol, this bot is full of bugs about when to make moves 2010-11-04T07:34:10 antimatroid1: what's the IP to? Your own testing server? 2010-11-04T07:34:19 using retard 2010-11-04T07:34:30 72.44.46.68 tcp ip 2010-11-04T07:35:08 *** narnach has quit IRC (Read error: Operation timed out) 2010-11-04T07:35:31 d0ugal: it's the main tcp server that everyone uses... run by dhartmei 2010-11-04T07:35:52 oh yeah, i should read more 2010-11-04T07:36:36 lol, retard was my username on there too, not a reference to you ;) 2010-11-04T07:38:31 *** genericbob has joined #aichallenge 2010-11-04T07:40:30 *** narnach has joined #aichallenge 2010-11-04T07:41:24 yeah 7 turn win 2010-11-04T07:42:26 make that 9 2010-11-04T07:49:08 antimatroid1: lol, np 2010-11-04T07:49:39 sigh: Don't think I know about that server :) but nevermind. 2010-11-04T07:49:50 @tcp 2010-11-04T07:49:51 antimatroid1: tcp = (#1) http://www.benzedrine.cx/planetwars/, or (#2) http://72.44.46.68/ 2010-11-04T07:49:52 so, everyone - me 2010-11-04T07:50:00 ah 2010-11-04T07:50:21 thanks 2010-11-04T07:50:32 if you're using this visualiser you can do it straight from inside the app 2010-11-04T07:50:32 http://ai-contest.com/forum/viewtopic.php?f=18&t=751 2010-11-04T07:52:13 it doesn't matter where you go, bocsimacko is miles ahead 2010-11-04T07:55:16 *** boegel has joined #aichallenge 2010-11-04T08:00:24 yeah, i got one against testingFromPiter 2010-11-04T08:00:28 do we know who that is? 2010-11-04T08:07:30 *** LEXASOFT has quit IRC (Quit: Page closed) 2010-11-04T08:14:58 Top 10 players: bocsimacko(4287), bix0r4ever(4103), jimrogerz(4102), george(4083), dmj111(4050), shangas(4026), GreenTea(4012), Accoun(3967), Hazard(3957), felixcoto(3943) 2010-11-04T08:16:06 http://72.44.46.68/canvas?game_id=851614 2010-11-04T08:16:14 that was going to be exciting till i sent an invalid order haa 2010-11-04T08:24:24 *** Blkt has joined #aichallenge 2010-11-04T08:25:08 Just having a look at the tcp server now 2010-11-04T08:25:20 I clearly missed something though, where do I get a username and password for it? 2010-11-04T08:28:44 *** Naktibalda has joined #aichallenge 2010-11-04T08:29:46 you just make them up 2010-11-04T08:29:53 if you want a new elo score, use a new one 2010-11-04T08:30:08 and you set the password on first use, usernames are freed after the last game was over 24 hours ago 2010-11-04T08:30:57 *** mceier has joined #aichallenge 2010-11-04T08:31:50 yup, usernames are free-for-all 2010-11-04T08:32:03 first come, first serve :P 2010-11-04T08:32:25 *** lavalamp_ has quit IRC (Quit: Page closed) 2010-11-04T08:39:16 *** boegel has quit IRC (Ping timeout: 276 seconds) 2010-11-04T08:48:35 obligatory shampoo keeps throwing matches 2010-11-04T08:49:23 http://72.44.46.68/canvas?game_id=851953 2010-11-04T08:49:24 example 2010-11-04T08:50:58 *** boegel has joined #aichallenge 2010-11-04T08:57:49 i hate dimwit, right before it loses it sends it ships to the furthest planet 2010-11-04T08:58:04 ah fuck, then he claimed it and sent them back again 2010-11-04T09:00:40 oh crap, how'd i let that happen 2010-11-04T09:01:15 might have been beyond your horizon 2010-11-04T09:01:26 http://72.44.46.68/canvas?game_id=852077 2010-11-04T09:01:41 terrible 2010-11-04T09:02:05 it got too cocky 2010-11-04T09:03:08 plus it's not sending fleets when wanted atm, but rather when "adding them" which results in quite a few stupid moves 2010-11-04T09:04:29 *** Appleman1234 has quit IRC (Ping timeout: 252 seconds) 2010-11-04T09:05:57 *** superflit has quit IRC (Quit: superflit) 2010-11-04T09:06:25 *** sigh has quit IRC (Remote host closed the connection) 2010-11-04T09:06:32 E323 around? 2010-11-04T09:07:30 at least use the full nickname :) 2010-11-04T09:08:29 heh thats what his bots run on tcp as :) 2010-11-04T09:08:31 E323! 2010-11-04T09:10:45 mega1 start up your bot on tcp 2010-11-04T09:10:58 I'm about to 2010-11-04T09:11:08 0.50 ? 2010-11-04T09:11:11 :p 2010-11-04T09:11:15 are you there yet 2010-11-04T09:13:12 here we go 2010-11-04T09:13:17 may my ISP be with me 2010-11-04T09:13:21 leave it a while, i'm putting a new one up soon 2010-11-04T09:13:24 you will still own me though 2010-11-04T09:13:36 *** phreeza has joined #aichallenge 2010-11-04T09:14:11 once i get it to do what i want without invalid ordering 2010-11-04T09:14:11 who is testingBot is that you antimatroid? 2010-11-04T09:14:47 nope, i'm retard 2010-11-04T09:14:57 hmm 2010-11-04T09:14:58 Top 10 players: bocsimacko(4280), george(4091), bix0r4ever(4090), jimrogerz(4067), dmj111(4050), GreenTea(4017), shangas(4003), Accoun(3969), Hazard(3946), felixcoto(3942) 2010-11-04T09:15:03 gah, i don't know why this is breaking, hmmm 2010-11-04T09:18:07 hmmm, not quite doing what i want, but i think i can get it to run without crashing or making stupid moves at neutrals 2010-11-04T09:18:30 *** perror has joined #aichallenge 2010-11-04T09:18:39 mega1: damn you aha 2010-11-04T09:18:49 antimatroid1: ? 2010-11-04T09:18:54 66 moves 2010-11-04T09:19:22 that was faint praise 2010-11-04T09:19:22 *** Rubicon-|-Cross has quit IRC (Quit: Leaving) 2010-11-04T09:19:56 i am in aww at your bot :P 2010-11-04T09:19:59 how's that for praise 2010-11-04T09:20:12 awe* same thing 2010-11-04T09:20:51 i want to play you on map 7 2010-11-04T09:21:04 it's my favourite 2010-11-04T09:21:28 that's pretty aggressive 2010-11-04T09:21:39 combative, I should say 2010-11-04T09:21:57 i just put in a new move selection, i'm working things out 2010-11-04T09:22:03 i'm trying to make it more defensive :P 2010-11-04T09:26:27 *** criis has left #aichallenge 2010-11-04T09:27:39 *** wh1teside has quit IRC (Read error: Connection reset by peer) 2010-11-04T09:28:56 *** wh1teside has joined #aichallenge 2010-11-04T09:29:54 mega1's bot is making me cry 2010-11-04T09:30:25 b0.48 is just too aggressive by the look of it 2010-11-04T09:30:36 my bot is too expansive atm 2010-11-04T09:31:03 hmm 2010-11-04T09:31:09 my but is doing some very dumb stuff 2010-11-04T09:31:56 err *bot 2010-11-04T09:32:48 IS NULL condition must be in WHERE 2010-11-04T09:32:57 wrong channel 2010-11-04T09:33:19 SQL 2010-11-04T09:33:21 !! 2010-11-04T09:34:12 yes, some exclamation marks were missing 2010-11-04T09:34:41 hmm 2010-11-04T09:35:07 btw what are you guys using as a metric to decide whether its safe to acquire a neutral, as in you have enough ships to acquire the neutral and not get wiped by the enemy 2010-11-04T09:35:16 i think my metric is off 2010-11-04T09:35:41 its either too conservative 2010-11-04T09:35:54 or too aggresive 2010-11-04T09:36:38 *** krokkrok has joined #aichallenge 2010-11-04T09:37:14 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716]) 2010-11-04T09:38:10 mine is non existent :P, althouggh i just switched to moves being to a destination without final times 2010-11-04T09:39:32 *** krokkrok has quit IRC (Client Quit) 2010-11-04T09:39:52 ah yours minmax so i guess as long as it sees it as an advantage it will take it 2010-11-04T09:40:08 it's not minimax atm 2010-11-04T09:40:12 strictly a decision tree 2010-11-04T09:42:39 *** Utkarsh has quit IRC (Ping timeout: 255 seconds) 2010-11-04T10:02:32 *** JamesMG has joined #aichallenge 2010-11-04T10:03:56 *** Azrathud has joined #aichallenge 2010-11-04T10:08:18 ok maybe someone here can help me put with this. My brain isn't working right it seems: 2010-11-04T10:08:48 this problem is surely equivalent to a standard CS problem, i'm trying to figure out which 2010-11-04T10:08:58 i superimpose 2 maps 2010-11-04T10:09:24 and i try to match up the planets from the 2 maps, so that the sum of the distances is minimal 2010-11-04T10:10:03 exhaustive search is too big, and greedy seems to screw up a lot 2010-11-04T10:10:12 *** superflit has joined #aichallenge 2010-11-04T10:10:17 any ideas? 2010-11-04T10:10:31 hmmm, some kind of regression maybe? 2010-11-04T10:10:35 to minimise dist 2010-11-04T10:10:56 regression wrt what? 2010-11-04T10:11:01 to clarify 2010-11-04T10:11:07 yeah, i'm thinking other stuff now :P 2010-11-04T10:11:10 i don't want to move about the map 2010-11-04T10:11:17 trying to work out how to do minimal spanning somehow 2010-11-04T10:11:18 just find the pairings 2010-11-04T10:11:38 one planet from map A to exactly one from map B 2010-11-04T10:12:14 bounded or not? 2010-11-04T10:12:40 yeah i think that'd work 2010-11-04T10:12:40 JamesMG: are you asking me? 2010-11-04T10:12:48 yeah phreeza 2010-11-04T10:12:54 sort all edges between two maps by distance 2010-11-04T10:13:05 add in new pairings that are new joinings 2010-11-04T10:13:07 a1 and a2 both have b2 as the closest planet? 2010-11-04T10:13:24 JamesMG: yea thats exactly the problem 2010-11-04T10:13:26 that should be fairly efficient 2010-11-04T10:13:31 thats why k-NN doesn't work 2010-11-04T10:13:32 *** yasith has quit IRC (Ping timeout: 265 seconds) 2010-11-04T10:13:41 phreeza: does that above work? 2010-11-04T10:13:56 kind of like finding a minimum spanning tree, but not 2010-11-04T10:14:14 antimatroid1: i think that is like my greedy algo... 2010-11-04T10:14:34 it works quite well in many cases, but then sometimnes it screws up 2010-11-04T10:14:35 ah, okay 2010-11-04T10:14:39 when there are clusters 2010-11-04T10:14:49 hmmmm 2010-11-04T10:14:56 cool idea :) 2010-11-04T10:14:59 Top 10 players: bocsimacko(4271), george(4082), bix0r4ever(4080), jimrogerz(4059), dmj111(4039), GreenTea(4010), shangas(3987), Accoun(3958), Raschi(3942), Hazard(3935) 2010-11-04T10:15:02 *** Naktibalda has joined #aichallenge 2010-11-04T10:15:05 JamesMG: what do you mean by bounded? 2010-11-04T10:15:15 can each planet be used only once? 2010-11-04T10:15:21 yes 2010-11-04T10:15:24 ah 2010-11-04T10:15:24 i think 2010-11-04T10:15:24 yes 2010-11-04T10:15:30 ok 2010-11-04T10:15:35 i want to have a mapping 2010-11-04T10:15:43 bijective? is that the word? 2010-11-04T10:15:50 bijective is a word 2010-11-04T10:16:06 can you give an example where my/your way would break? 2010-11-04T10:16:23 ok give me a minute 2010-11-04T10:18:59 does it happen when there exists more than one bijective map and you don't get the same both ways? 2010-11-04T10:19:12 i would imagine you could fix that by enforcing planet numbering? 2010-11-04T10:19:23 min sum of both planet numbers added or something 2010-11-04T10:20:02 hmmm, i guess that can still tie break 2010-11-04T10:20:03 ummmm... 2010-11-04T10:20:44 you could use map 1 as that which minimises average distance, but even that can tie 2010-11-04T10:22:21 erm... i think my code is wrong... but i am thinking of an artificial example... just a second :D 2010-11-04T10:26:38 felixcoto is testing a new bot 2010-11-04T10:27:03 antimatroid1: ok i drew a sketch :) uploading photo 2010-11-04T10:27:07 * antimatroid1 takes a peek 2010-11-04T10:27:26 mega1: you beat him anyway :P 2010-11-04T10:27:32 that's the old one 2010-11-04T10:27:37 antimatroid1: http://twitpic.com/33twm8 2010-11-04T10:27:47 oh, yes 2010-11-04T10:28:28 phreeza: what do you want the mapping to be? 2010-11-04T10:28:44 damn should have added numbers :) 2010-11-04T10:28:55 do it in paint? :P 2010-11-04T10:28:56 so let the blue ones be a1..a3 2010-11-04T10:29:00 from top to bottom 2010-11-04T10:29:01 yep 2010-11-04T10:29:14 and b1 top left 2010-11-04T10:29:17 b2 right 2010-11-04T10:29:20 b3 bottom 2010-11-04T10:29:27 terrible :P 2010-11-04T10:29:30 b should be blue 2010-11-04T10:29:36 doh! 2010-11-04T10:29:38 :D 2010-11-04T10:30:00 go on, with your naming convention :P 2010-11-04T10:30:28 anyway, greedy will map a1->b1 a2->b3 a3->b2 2010-11-04T10:31:05 yeah 2010-11-04T10:31:06 but it should be an->bn 2010-11-04T10:31:20 hmmmm 2010-11-04T10:31:27 would dp do it? 2010-11-04T10:31:33 dp == dynamic programming 2010-11-04T10:31:56 that might be your best bet 2010-11-04T10:32:03 probably somehow... but i don't know how to adapt the algos i know to that problem 2010-11-04T10:32:05 mega1: any suggestions 2010-11-04T10:32:12 should read up on that 2010-11-04T10:32:25 is it the planet mapping problem still? 2010-11-04T10:32:29 thats what i used for my knapsack bot in the early days :) 2010-11-04T10:32:34 yes 2010-11-04T10:32:42 *** davidjliu has joined #aichallenge 2010-11-04T10:32:42 just use back recursion, for planet 1 find best planets, then for planet 2 find best with planet 1's best and then work your way through 2010-11-04T10:33:10 tbh my cs education is fairly poor 2010-11-04T10:33:10 my gut feeling is that it's equivalent to the traveling salesman 2010-11-04T10:33:33 in complexity, at least :-) 2010-11-04T10:33:58 that would be bad news :) 2010-11-04T10:34:01 :( 2010-11-04T10:34:19 you have 23 planets, you need to do it how often? 2010-11-04T10:34:20 you only need a very solution, right? 2010-11-04T10:34:25 *very good 2010-11-04T10:34:44 yes 2010-11-04T10:35:26 then perturb the greedy solution 2010-11-04T10:35:32 simulated annealing comes to mind 2010-11-04T10:35:53 ah... thats an interesting idea 2010-11-04T10:36:33 the word "perturbation" just reminds me of calculus of variations 2010-11-04T10:36:49 given every damn proof was like, okay now perturb our functional 2010-11-04T10:36:58 antimatroid1: physicist? 2010-11-04T10:37:09 cs, economics and maths student 2010-11-04T10:37:13 just finishing up my undergrad now 2010-11-04T10:37:27 though it might be that if there are local minima in these clusters 2010-11-04T10:37:34 ah ok 2010-11-04T10:37:45 i am a phys student, thats why i ask ;) 2010-11-04T10:37:56 i have quite a few of them as friends :P 2010-11-04T10:38:09 cool 2010-11-04T10:38:37 the maths and physics departments at my uni are pretty much one in the same 2010-11-04T10:38:45 is this to generalize in strategy space? 2010-11-04T10:38:58 similarity function over maps? 2010-11-04T10:39:01 *** Utkarsh has joined #aichallenge 2010-11-04T10:39:01 yea kinda 2010-11-04T10:39:10 phreeza: was it you that was talking about trying to set up equivalence classes between maps? 2010-11-04T10:39:12 i want to have a ddatabase of precedents 2010-11-04T10:39:19 yes 2010-11-04T10:39:41 so to do that i want to do clustering on the maps 2010-11-04T10:39:42 any love? 2010-11-04T10:39:58 its just a bunch of thought in my head really 2010-11-04T10:40:07 started actually coding last night 2010-11-04T10:40:15 nothing running yet 2010-11-04T10:40:27 but i'm looking forward to datamining the game logs 2010-11-04T10:40:28 :) 2010-11-04T10:41:01 i wouldn't have a clue how to do anything like that 2010-11-04T10:41:28 i read a couple of papers on case based reasoning 2010-11-04T10:41:33 and case based planning 2010-11-04T10:41:46 plus the tiniest change of numbers in this can change things completely 2010-11-04T10:42:10 yea so the idea is to have a feature vector that accounts for that 2010-11-04T10:42:19 won't be counting actual number of ships 2010-11-04T10:42:22 but stuff like 2010-11-04T10:42:35 is fleet a big enough to take that planet 2010-11-04T10:42:36 or not 2010-11-04T10:42:48 the responses also 2010-11-04T10:42:55 more of an abstract nature 2010-11-04T10:44:16 the thing is even in the best case this will only mimic the top bots 2010-11-04T10:44:24 don't see how it can get better than them ;) 2010-11-04T10:44:34 even that is unlikely ;) 2010-11-04T10:45:14 i'll be happy when i can consistently pound mega1 2010-11-04T10:45:28 hehe who wouldn't 2010-11-04T10:45:45 i have 3 weeks of nothing but ai before the contest ends after tuesday :) 2010-11-04T10:46:07 :/ i have plkenty of other stuff to work on 2010-11-04T10:48:59 ah i think i found somethiung on wikipedia 2010-11-04T10:49:01 http://en.wikipedia.org/wiki/Assignment_problem 2010-11-04T10:49:25 mega1: one solution is called Hungarian Algorithm 2010-11-04T10:49:56 heh, sorry to dissapoint 2010-11-04T10:50:59 n^3 can be steep 2010-11-04T10:51:48 i'm happy its polynomial :) 2010-11-04T10:52:56 oO solved in 1890 2010-11-04T10:53:00 by Jacobi 2010-11-04T10:53:11 standing on the shoulders of giants ;) 2010-11-04T10:54:39 lol, i don't even want to work out how bad the complexity is for my move selection 2010-11-04T10:54:50 it'd be at least n^4 i think 2010-11-04T10:54:56 maybe 2010-11-04T10:55:00 not sure it could handle that? 2010-11-04T10:55:05 maybe it is just n^3 2010-11-04T10:55:10 i am terrible with that kind of thing :P 2010-11-04T10:55:18 Knuths implementation :D 2010-11-04T10:55:20 http://reptar.uta.edu/NOTES5311/hungarian.c 2010-11-04T11:07:16 mega1: i'm having your redistribution issues, i'm going to take that as a good thing 2010-11-04T11:09:55 *** superflit has quit IRC (Quit: superflit) 2010-11-04T11:13:11 mega1: do you assume the enemy will make the same move you would or is your position purely evaluated 2010-11-04T11:13:22 i'm having a hard time figuring out how to attribute enemy possibilities 2010-11-04T11:14:59 Top 10 players: bocsimacko(4264), george(4087), bix0r4ever(4079), jimrogerz(4033), dmj111(4030), GreenTea(4018), shangas(3971), Accoun(3959), Raschi(3939), felixcoto(3937) 2010-11-04T11:15:13 demerzel_: I have a 1 ply and a nash equilibrium 2 ply search 2010-11-04T11:15:19 and 2 ply minimax too 2010-11-04T11:15:21 really 1 ply? 2010-11-04T11:15:33 the 1 ply is the best 2010-11-04T11:15:39 mega1: ever have luck with simultaneous games? 2010-11-04T11:15:44 do you just let all 3 run and pick the best? 2010-11-04T11:15:49 how do you eval the best in that case 2010-11-04T11:15:58 i am almost to the point of having a stable 1 ply minimax 2010-11-04T11:16:14 antimatroid1: not much, it's much slower and slightly worse 2010-11-04T11:16:23 I retest it every now and then. 2010-11-04T11:16:39 antimatroid: what do you mean by simultaneous games? 2010-11-04T11:16:42 demerzel_: I run local tests of few hundred games 2010-11-04T11:16:59 demerzel_: tree approach generally drops the imperefect information part 2010-11-04T11:17:15 err can you explain that a little more 2010-11-04T11:17:15 leaving it a game with sequential moves, which it is not 2010-11-04T11:17:26 ... and mislead myself tremendously sometimes 2010-11-04T11:17:28 ohh you mean sequential vs simultaneous 2010-11-04T11:17:34 yes 2010-11-04T11:17:36 yes 2010-11-04T11:17:49 mega1: on what type of actions is the minimax search? 2010-11-04T11:17:59 i see no reason to make it simultaneous all the stuff i see about that adds several layers of complexity 2010-11-04T11:18:17 i guess you have to give the enemy a minimax move to make it even tho 2010-11-04T11:18:21 it's not that hard for 2 player zero sum games unless you want mixed strategy nash equilibria 2010-11-04T11:18:48 pure strategy nash equilibria is easy enough even with n players for normal games 2010-11-04T11:18:53 mixed strategies are always where it hurts 2010-11-04T11:19:13 i am slowly coming to what i believe is a pure strategy minimax 2010-11-04T11:19:18 *** Olathe has joined #aichallenge 2010-11-04T11:19:19 i am still struggling with all these game theory terms 2010-11-04T11:20:49 *** Azrathud has quit IRC (Ping timeout: 260 seconds) 2010-11-04T11:20:54 if anyone knows a good 'game theory for retards who barely speak english' page that would help 2010-11-04T11:21:12 Error323: composite orders 2010-11-04T11:21:41 mega1: do you specify a completion time for moves, or do it "into the future" per planet? 2010-11-04T11:21:45 i recently switched to the later 2010-11-04T11:22:02 not really working very well for me yet though 2010-11-04T11:22:06 i see 2010-11-04T11:22:09 I have a concrete time for future moves 2010-11-04T11:22:32 *** lavalamp_ has joined #aichallenge 2010-11-04T11:22:40 *** yasith has joined #aichallenge 2010-11-04T11:22:51 i have a concrete time for fleets inside moves, but each move is now for a planet from time 1 into the future 2010-11-04T11:23:27 well i wasn't going to ask but you guys are so far ahead 2010-11-04T11:23:38 what do you count as the next move in a sequence? arrival or same turn? 2010-11-04T11:23:48 *same turn with results 2010-11-04T11:24:33 i was toying with the idea of creating a tree based on sequential moves after arrival 2010-11-04T11:24:44 but then i either have to have n trees or have some way to know if my current position is worse 2010-11-04T11:25:11 for a move i will use ships that i own at any planet at any time in the future, so i make one move, update and then make another move based off that information 2010-11-04T11:25:15 if that makes sense 2010-11-04T11:25:21 so same turn 2010-11-04T11:25:32 sort of not really 2010-11-04T11:25:35 it does all turns at once 2010-11-04T11:25:44 although i wont go past the max distance of the map 2010-11-04T11:26:10 so you use the natural depletion of resources as a progression 2010-11-04T11:26:20 ? 2010-11-04T11:26:43 ie the move could be this turn or 10 turns out 2010-11-04T11:26:56 if it's 10 out then it's because you don't have the resources now 2010-11-04T11:27:09 i scan through and each planet has the new amount of ships there at that time, that's what i can send from that planet at that time point 2010-11-04T11:27:10 not because you specified it's 10 out 2010-11-04T11:27:19 right? 2010-11-04T11:27:29 a fleet has a source, time sent and destination, whereas a move consists of many fleets but a single destination 2010-11-04T11:27:54 yeah, although i recently moved to a stack for things like that, so i add stuff later rather than earlier 2010-11-04T11:28:11 fuuuuuuuuck yeah, i fixed my failing against noobs problem: http://72.44.46.68/canvas?game_id=853497 2010-11-04T11:28:20 a stack as in a coding container or is that a game theory term too 2010-11-04T11:28:28 coding container 2010-11-04T11:28:33 just checking haha 2010-11-04T11:28:35 I use a stack for my orders also... 2010-11-04T11:28:37 i have lots of stacks/queues/priority queues 2010-11-04T11:28:42 my definition of minmax was quite different before this 2010-11-04T11:28:58 you can do minimax differently with simultaneous games 2010-11-04T11:30:10 hmm just trying to get a stable base game - it seems to do ok now 2010-11-04T11:30:19 *** aerique has quit IRC (Quit: ...) 2010-11-04T11:30:32 although it's way too aggressive and has no perception of if its winning or lose 2010-11-04T11:31:02 Error323: you only won that because shampoo has learned from my bot how to waste ships on neutrals... 2010-11-04T11:31:14 shampoo always has done that 2010-11-04T11:31:22 lavalamp_: no 2010-11-04T11:31:26 i swear his playstyle is like my old ideal 2010-11-04T11:31:26 i was losing 2010-11-04T11:31:38 ah well... my bot independantly discovered that then 2010-11-04T11:31:48 and by observing that, i took action :) 2010-11-04T11:31:55 ah ok 2010-11-04T11:32:21 as in a remembered observation or just a numerical advantage opening 2010-11-04T11:32:34 yeah on my list is to massively increase priority of taking neutrals if I'm behind in production 2010-11-04T11:32:36 the former would e impressive hehe 2010-11-04T11:32:46 if you look at turn 0 - 70 i have less planets 2010-11-04T11:32:49 and keep attacking 2010-11-04T11:33:06 then i observe that its not really working and i cap neutrals @ 75 2010-11-04T11:33:16 which win me the game 2010-11-04T11:34:33 *** O1athe has joined #aichallenge 2010-11-04T11:34:35 hm, after 93 you could have sniped the 66 ships neutral 2010-11-04T11:35:10 actually anytime after 84 2010-11-04T11:35:34 mmm 2010-11-04T11:35:36 good point 2010-11-04T11:36:08 instead of sending all those ships to the planet behind it... 2010-11-04T11:36:26 not that you need my help XP 2010-11-04T11:36:35 *** O1athe has quit IRC (Read error: Connection reset by peer) 2010-11-04T11:36:46 well but ur right 2010-11-04T11:36:54 *** superflit has joined #aichallenge 2010-11-04T11:37:35 *** Olathe has quit IRC (Ping timeout: 240 seconds) 2010-11-04T11:37:48 lavalamp_: well actually 2010-11-04T11:37:53 sniping would have failed there 2010-11-04T11:38:03 he would've reinforced it 2010-11-04T11:38:11 before i got there 2010-11-04T11:38:29 *** bhasker_ has quit IRC (Quit: bhasker_) 2010-11-04T11:41:17 woop something worked right the first time - total scores exactly correct haha 2010-11-04T11:41:19 Error323: Yeah I guess you're right... but it still makes more sense as a target than the planet behind it... 2010-11-04T11:44:49 *** p4p4 has joined #aichallenge 2010-11-04T11:50:07 *** Cyndre has joined #aichallenge 2010-11-04T11:50:23 lol i put up a new bot and it's completely retarded 2010-11-04T11:52:03 well i just found a loophole that allows a 0 cost move 2010-11-04T11:52:10 that's gonna be fun to track down 2010-11-04T11:53:41 oh nm i was sending exactly as many as were on the planet haha 2010-11-04T11:53:50 i don't own it yet we need 0 more ships! 2010-11-04T11:55:47 *** Itkovian has quit IRC (Quit: Itkovian) 2010-11-04T12:02:16 Man, I read that, run a test game batch, and find the same error 2010-11-04T12:02:41 sorry :) 2010-11-04T12:02:42 quick, stream enemyPlanet.GrowthRate() at the enemy planet! 2010-11-04T12:02:49 that'll teach 'em 2010-11-04T12:03:15 well now i have teh same problem with defense 2010-11-04T12:03:22 it's never happy because it doesn't get updated haha 2010-11-04T12:03:37 forgot to make that a parameterized update instead of a global one 2010-11-04T12:09:07 *** Rubicon-|-Cross has joined #aichallenge 2010-11-04T12:14:59 Top 10 players: bocsimacko(4273), george(4084), bix0r4ever(4071), jimrogerz(4036), dmj111(4035), GreenTea(4019), shangas(3983), Accoun(3961), Hazard(3944), Raschi(3940) 2010-11-04T12:28:25 *** boegel has quit IRC (Ping timeout: 255 seconds) 2010-11-04T12:30:22 oh my god my bot won a game 2010-11-04T12:30:29 *against some guy who capped 2 planets and sat there 2010-11-04T12:30:40 my bot's counter: cap 3 and sit there 2010-11-04T12:33:03 hey, whatever works, right? :) 2010-11-04T12:37:23 *** wh1teside_ has joined #aichallenge 2010-11-04T12:39:08 *** wh1teside has quit IRC (Read error: Connection reset by peer) 2010-11-04T12:41:59 *** MrG_ has joined #aichallenge 2010-11-04T12:46:17 @tcp 2010-11-04T12:46:18 lavalamp_: tcp = (#1) http://www.benzedrine.cx/planetwars/, or (#2) http://72.44.46.68/ 2010-11-04T12:48:29 http://72.44.46.68/canvas?game_id=853172 2010-11-04T12:48:32 what a great fight 2010-11-04T12:50:21 nice 2010-11-04T12:50:43 would be cool to have the updated viewer on the tcp server 2010-11-04T12:50:51 woop my bot is getting less retarded 2010-11-04T12:50:51 http://72.44.46.68/canvas?game_id=854172 2010-11-04T12:50:59 slowly but surely 2010-11-04T12:51:05 no dist code yet :( 2010-11-04T12:51:53 *** Migi32 has joined #aichallenge 2010-11-04T12:52:17 holy cow the lamp above my head fell down. Right next to me. :o 2010-11-04T12:52:27 what's the "awesomeness" about the new viewer? 2010-11-04T12:52:46 did it give you an insight of newtonian proportions? 2010-11-04T12:53:08 it plots growth rares too 2010-11-04T12:53:16 nope. It didn't fall on my head :( 2010-11-04T12:53:16 rates* 2010-11-04T12:53:31 *** bhasker has joined #aichallenge 2010-11-04T12:53:55 and the plots have a cursor to show you which turn is happening 2010-11-04T12:53:58 and speed controls 2010-11-04T12:54:16 yes. The cursor is pretty useful. 2010-11-04T12:54:29 Any plans on TCP adopting the new visualiser? 2010-11-04T12:54:42 ehhh 2010-11-04T12:54:49 i don't see all that 2010-11-04T12:54:55 @seen dhartmei 2010-11-04T12:54:55 phreeza: dhartmei was last seen in #aichallenge 5 weeks, 1 day, 0 hours, 10 minutes, and 17 seconds ago: project selection problem? 2010-11-04T12:54:56 is your version flash? 2010-11-04T12:55:17 oh. Strange that he still runs his bot on TCP even if he's not around anymore 2010-11-04T12:55:41 -he runs the tcp server i think 2010-11-04T12:55:54 yea I know 2010-11-04T12:56:21 Error323: here for example, you don't see the changed? 2010-11-04T12:56:26 changes 2010-11-04T12:56:27 http://ai-contest.com/visualizer.php?game_id=6612786 2010-11-04T12:56:33 Error323, there is no flash visualiser. Only java and html5 2010-11-04T12:57:13 i see a + and - that don't do anything here, and a bar below and above the graph 2010-11-04T12:57:21 no growthrate or sth 2010-11-04T12:57:31 oh holy shit 2010-11-04T12:57:33 now i do 2010-11-04T12:57:34 o_O 2010-11-04T12:57:44 this is really nice yeah 2010-11-04T12:58:43 *** Mekanik has joined #aichallenge 2010-11-04T12:58:52 whoah its actually a slider 2010-11-04T12:58:55 freaking awesome 2010-11-04T12:59:02 yeah :) 2010-11-04T13:00:10 and as I log on to the main site to test out the slider (because I didn't know that, ty Error323 :D) I see this: 2010-11-04T13:00:11 http://ai-contest.com/profile.php?user_id=8565 2010-11-04T13:00:25 scary. 2010-11-04T13:00:39 jeeep 2010-11-04T13:00:47 its a pretty sight 2010-11-04T13:00:49 when its ur bot 2010-11-04T13:01:04 but if it's not... 2010-11-04T13:01:40 jesus, on TCP too: 2010-11-04T13:01:40 http://72.44.46.68/getplayer?player=bocsimacko-b0.47 2010-11-04T13:01:47 http://72.44.46.68/getplayer?player=bocsimacko-b0.48 2010-11-04T13:01:56 :o 2010-11-04T13:02:01 did he solve PlanetWars? 2010-11-04T13:02:26 i think he came pretty close to that yea 2010-11-04T13:02:33 given the methods he's using 2010-11-04T13:02:39 not quite :) http://ai-contest.com/visualizer.php?game_id=6436339 2010-11-04T13:03:28 true, I win sometimes aswell :D 2010-11-04T13:03:32 phreeza, is that his latest submission? 2010-11-04T13:03:40 but its truely mindblowing =) 2010-11-04T13:05:33 maybe those games are just him having bad luck on a rock-paper-scissor type situation 2010-11-04T13:05:57 though I don't know why he counter-attacked you in that game, phreeza. He was ahead. 2010-11-04T13:07:44 *** Bobng has joined #aichallenge 2010-11-04T13:08:09 not my bot hehe 2010-11-04T13:08:15 just a submission i found 2010-11-04T13:14:05 does his bot use the full 1 second when you play him on TCP? 2010-11-04T13:14:59 Top 10 players: bocsimacko(4268), george(4089), bix0r4ever(4067), dmj111(4049), jimrogerz(4031), GreenTea(4010), shangas(3982), Accoun(3968), Raschi(3950), felixcoto(3933) 2010-11-04T13:16:37 Migi32: takes about 0.1s per turn 2010-11-04T13:16:52 *** JamesGecko_ has joined #aichallenge 2010-11-04T13:17:41 oh. So no tree-searches or anything. mega1, it's mostly heuristics then, like 99% of us? 2010-11-04T13:17:57 Migi32: 1 ply search 2010-11-04T13:18:12 ok. Thanks for sharing :) 2010-11-04T13:18:40 I'm testing nash equilibrium again, but it's not stronger 2010-11-04T13:18:47 althought a whole lot slower 2010-11-04T13:20:46 protip- when putting in distribution code make sure the planet it's trying to send to is yours 2010-11-04T13:20:46 mega1: nash? 2010-11-04T13:20:58 yes 2010-11-04T13:21:03 * raedwulf googles 2010-11-04T13:21:45 I've been experimenting with Nash equilibria in my bot a lot myself. But haven't really found a good way to use it either; 2010-11-04T13:21:50 oh ok 2010-11-04T13:23:01 nash requires a level of comprehending the gamestate that i am not capable of atm 2010-11-04T13:23:25 how can you even lock it down to that level 2010-11-04T13:25:24 mega1, what are you using to calculate the Nash equilibria? Linear programming? 2010-11-04T13:25:38 someone posted a link the other day 2010-11-04T13:25:57 http://code.activestate.com/recipes/496825/ 2010-11-04T13:27:27 mega1, ah ok it returns an approximation. I wasted a lot of time with linear programming and simplex to get exact numbers. 2010-11-04T13:27:48 no need to 2010-11-04T13:29:24 yea that code snippet is like at least 20 times shorter than all the code I have for it 2010-11-04T13:34:28 Oh, nice link, that's shorter and better than what I had 2010-11-04T13:34:37 *** wh1teside has joined #aichallenge 2010-11-04T13:34:45 does redistribution work for you guys 2010-11-04T13:34:49 i can never get it to work properly 2010-11-04T13:35:01 i always end up in a worse situation when i try to redistribute ships 2010-11-04T13:35:07 mine's broken atm 2010-11-04T13:35:15 it does here 2010-11-04T13:35:28 you mean staging bhasker ? 2010-11-04T13:35:33 no 2010-11-04T13:35:41 oh 2010-11-04T13:35:43 what do you mean? 2010-11-04T13:35:43 i mean moving ships to planets that may get sniped etc 2010-11-04T13:35:56 like figuring out vulernable planets and moving ships there preemptively 2010-11-04T13:36:01 vulnerable* 2010-11-04T13:36:03 *** wh1teside_ has quit IRC (Ping timeout: 265 seconds) 2010-11-04T13:36:03 *** narnach has quit IRC (Quit: Leaving.) 2010-11-04T13:36:03 ahh 2010-11-04T13:36:39 oh I call that fortification 2010-11-04T13:37:03 ...and mine's broken atm... 2010-11-04T13:37:11 yea whenever i put that code in 2010-11-04T13:37:19 i just end up with bunch of ships moving back and forth:-\ 2010-11-04T13:37:26 and doing nothing really 2010-11-04T13:37:29 haha 2010-11-04T13:37:38 mine doesn't do that, it just crashes 2010-11-04T13:38:17 heh 2010-11-04T13:38:29 i though i had the minimal ships required to hold the planet figured out 2010-11-04T13:38:31 do you look for ships on your target planet before looking off-planet? 2010-11-04T13:38:39 as in? 2010-11-04T13:38:44 *** narnach has joined #aichallenge 2010-11-04T13:39:14 how are you guys figuring out minimal ships required 2010-11-04T13:39:21 *** narnach has quit IRC (Client Quit) 2010-11-04T13:39:23 Whenever I want to send x ships to planet i by turn t, I always start looking on planet i at turn t and then go backwards in time 2010-11-04T13:39:26 all my current calculations seem to be always off by some 2010-11-04T13:39:34 hmm 2010-11-04T13:39:35 interesting 2010-11-04T13:39:51 when you say backwards in time 2010-11-04T13:39:54 what do you do 2010-11-04T13:40:03 when I get to the current turn, then I search adjacent planets in the same way 2010-11-04T13:40:26 hmm not sure i understand 2010-11-04T13:40:30 how many newly available ships are there this turn? reserve them and go to the previous turn 2010-11-04T13:40:37 hmm 2010-11-04T13:40:47 so when you say reserve them you add the as available? 2010-11-04T13:40:54 so you launch less from your initial planet? 2010-11-04T13:41:25 *** narnach has joined #aichallenge 2010-11-04T13:41:27 I keep track of how many ships are "held" at each planet at each turn 2010-11-04T13:41:40 hmm i don't retain state across turns 2010-11-04T13:41:50 nor do i remember what i did last turn 2010-11-04T13:41:50 each order has an array of held ships associated with it 2010-11-04T13:42:02 nor do I, I'm talking about in the future here 2010-11-04T13:42:22 hmm 2010-11-04T13:44:03 here's the relevant function: http://pastebin.com/A9zJNpGX 2010-11-04T13:44:22 that code has bugs in it which I fixed this morning but maybe it will get the idea across 2010-11-04T13:45:16 what language is that Go? 2010-11-04T13:45:17 The key is having a working NewlyAvailibleShips metric 2010-11-04T13:45:19 yeah, go 2010-11-04T13:45:26 ah k 2010-11-04T13:46:29 about to see how my forwarding code went haha 2010-11-04T13:46:40 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716]) 2010-11-04T13:51:15 *** MrG_ has joined #aichallenge 2010-11-04T13:57:44 *** Sarin has joined #aichallenge 2010-11-04T14:01:44 *** MrG_ has quit IRC (Quit: Page closed) 2010-11-04T14:05:56 *** JamesGecko_ has quit IRC (Ping timeout: 245 seconds) 2010-11-04T14:14:24 *** cfaftw has joined #aichallenge 2010-11-04T14:15:00 Top 10 players: bocsimacko(4249), george(4072), dmj111(4039), GreenTea(4019), Accoun(3995), shangas(3992), jimrogerz(3981), Raschi(3974), felixcoto(3948), Hazard(3930) 2010-11-04T14:19:42 *** smellyhippy has quit IRC (Ping timeout: 250 seconds) 2010-11-04T14:20:41 *** smellyhippy has joined #aichallenge 2010-11-04T14:20:42 *** smellyhippy has joined #aichallenge 2010-11-04T14:30:35 *** cfaftw has quit IRC (Quit: Leaving) 2010-11-04T14:33:11 *** p4p4 has quit IRC (Quit: ChatZilla 0.9.84 [SeaMonkey 2.0a3/20090223135443]) 2010-11-04T14:41:47 does anyone have any links to payoff matrices as examples instead of just using prisoner's dilemna like all the ones i've seen 2010-11-04T14:42:49 bummer, my C# bot has started crashing every game on the official server 2010-11-04T14:45:05 JamesMG: hmm, link to your profile? 2010-11-04T14:45:33 looking at the semaphores on the main server at least they look to be clear 2010-11-04T14:46:28 http://ai-contest.com/profile.php?user_id=10527 2010-11-04T14:46:56 has crashed the last 3 2010-11-04T14:48:12 doesn't look like it's the semaphore problem anyway. That crashes on the first turn I'm pretty sure 2010-11-04T14:48:29 yeah, it used to 2010-11-04T14:48:50 I'm wondering if it's actually timing out, since that code has done about 450 clean games on the TCP server 2010-11-04T14:48:57 but that's more lenient time wise 2010-11-04T14:49:08 wow 2010-11-04T14:49:19 i resubmitted my bot last night when i realized i forgot to turn off logging 2010-11-04T14:49:28 now it is 5-0 instead of 1-8 2010-11-04T14:49:35 yes, quite a bit more lenient 2010-11-04T14:50:06 *** boegel has joined #aichallenge 2010-11-04T14:50:13 there is another loss by timeout as well, second loss after resubmission 2010-11-04T14:50:54 ah, didn't notice that one 2010-11-04T14:51:06 ah, one of those games i should have lost 2010-11-04T14:51:40 and one was opponent didn't move on turn 1 2010-11-04T14:55:05 code runs in about .1 of a second locally 2010-11-04T14:56:00 can you try the specific situations seen in the timeouts on the main server? 2010-11-04T14:56:19 yeah, I'll throw those at it and see if it's the bot 2010-11-04T14:56:42 JamesMG: the spz.nofate viz has a translate to state 2010-11-04T14:56:58 ie it takes the replay string and there's a button down at the bottom to spit out what the bot saw 2010-11-04T14:57:15 just noticed that yesterday very useful 2010-11-04T14:57:26 awesome, thanks 2010-11-04T14:58:28 i can't figure out why my bot doesn't defend all the time 2010-11-04T14:58:33 it's pissing me off lol 2010-11-04T14:58:42 it knows when and how much 2010-11-04T14:58:43 just doesn't do it 2010-11-04T14:58:48 i even rigged eval to always give positive 2010-11-04T14:59:59 some orders getting popped off your stack prematurely? 2010-11-04T15:00:34 not sure i'm turning on some super logging to try and track it down 2010-11-04T15:00:42 it says 'hey i need 5 ships in 8 turns and yep i got em' 2010-11-04T15:00:44 then doesn't do it lol 2010-11-04T15:02:18 is something higher priority taking your ships first? 2010-11-04T15:02:25 only done one so far janzert, but it spits orders right back at me 2010-11-04T15:02:45 let me try it on mono 2010-11-04T15:04:44 well atm it's a pure value consideration 2010-11-04T15:04:49 and defense would naturally have the highest value 2010-11-04T15:05:05 less turns plus i dont' have to consider enemy fleet 2010-11-04T15:05:14 and i doubled it anyway lol 2010-11-04T15:05:38 i probably just missed a check somewhere - bot is almost ok again :) 2010-11-04T15:05:52 have you checked that you don't get NaN's out of your eval func? 2010-11-04T15:06:20 yep - i don't, evals are a struct that's pre-cleared 2010-11-04T15:06:26 I had those for a while 2010-11-04T15:06:32 NaNs? 2010-11-04T15:06:34 i thought c would just crash 2010-11-04T15:06:56 floating point values get set to NaN if you do something invalid 2010-11-04T15:07:09 JamesMG: could you link the spz visualizer of one of those games, I might be able to try running it on the main server and see if it reproduces there 2010-11-04T15:07:34 http://www.mail-archive.com/mono-bugs@lists.ximian.com/msg25959.html 2010-11-04T15:07:39 this is my current theory janzert 2010-11-04T15:07:43 ah i found it 2010-11-04T15:07:46 since my reader is using ReadLine 2010-11-04T15:07:53 if (planet is mine and planet is enemy) { consider defense } 2010-11-04T15:07:53 oops 2010-11-04T15:07:54 unless C has code to detect and crash on those, which I don't think it does 2010-11-04T15:07:56 http://spz.nofate.com/planetwars/visualizer/?game_id=6615329&server=Official 2010-11-04T15:07:58 lol 2010-11-04T15:08:22 no, they just stay nan 2010-11-04T15:08:33 I've got it crashing if I paste the state in using mono now 2010-11-04T15:08:38 it's just an overflow 2010-11-04T15:08:39 yeah that's what I thought 2010-11-04T15:08:41 ahh, ok 2010-11-04T15:08:51 guess I need to get away from ReadLine 2010-11-04T15:08:59 great yet another mono bug :} 2010-11-04T15:09:04 yeah no kidding :( 2010-11-04T15:09:09 once you get a NaN it pollutes your entire chain of calculations... 2010-11-04T15:09:12 might be useful to post it on the forums as well if you don't mind 2010-11-04T15:09:17 sure thing 2010-11-04T15:10:00 doing any sort of arithmetic will get you one of three values: nan, +inf, -inf 2010-11-04T15:10:13 *any sort of arithmetic with nan 2010-11-04T15:10:17 you want it in your announcement thread or should I put it in technical issues? 2010-11-04T15:10:42 there are actually different NaN values depending on what you did to get the NaN 2010-11-04T15:10:42 *** narnach has quit IRC (Ping timeout: 255 seconds) 2010-11-04T15:11:35 why did it overflow 2010-11-04T15:11:41 james 2010-11-04T15:11:50 mono bug 2010-11-04T15:11:57 ReadLine can't handle more than 1025 characters in Console apparently? 2010-11-04T15:12:00 in mono. 2010-11-04T15:12:20 i guess they use a static buffer of 1024? 2010-11-04T15:12:33 why would readline try to get that whole thing tho 2010-11-04T15:12:38 beats me 2010-11-04T15:12:44 because it's mono? :) 2010-11-04T15:13:08 its an old version of mono too :p 2010-11-04T15:13:08 lemme see what i was using 2010-11-04T15:13:20 oh you can fix that 2010-11-04T15:14:01 I'm just going to switch off of ReadLine, it's convenient but not necessary for this 2010-11-04T15:14:05 i can't find my code 2010-11-04T15:14:14 but basically just declare a binarystream with a buffer of say 4096 2010-11-04T15:14:20 and then point console.in to it 2010-11-04T15:14:49 try that and see if it still crashes 2010-11-04T15:14:57 I'm amazed I haven't hit this problem before 2010-11-04T15:15:00 Top 10 players: bocsimacko(4216), george(4081), dmj111(4038), GreenTea(4021), Accoun(3999), shangas(3993), Raschi(3978), felixcoto(3964), jimrogerz(3953), Hazard(3930) 2010-11-04T15:15:02 yea really 2010-11-04T15:15:13 *** narnach has joined #aichallenge 2010-11-04T15:16:49 either that or you need to set encoding specifically 2010-11-04T15:16:55 maybe it overran the buffer looking for \r\n 2010-11-04T15:17:04 if it's seriously 1024 tho that's a major issue 2010-11-04T15:21:45 *** boegel_ has joined #aichallenge 2010-11-04T15:25:31 *** boegel_ has quit IRC (Client Quit) 2010-11-04T15:27:14 *** perror has quit IRC (Quit: Bye all !) 2010-11-04T15:28:05 *** RainCT has joined #aichallenge 2010-11-04T15:30:13 *** yasith has quit IRC (Ping timeout: 276 seconds) 2010-11-04T15:30:13 OMG I JUST HAD A BRILLIANT IDEA FOR MY BOT!!! 2010-11-04T15:30:16 wrapping stdin with a bufferedstream seems to have fixed it 2010-11-04T15:30:24 oh, wait 2010-11-04T15:30:29 forget it 2010-11-04T15:30:36 lol 2010-11-04T15:30:42 now I'm curious boegel 2010-11-04T15:30:46 spill it :) 2010-11-04T15:31:06 I was just joking :) 2010-11-04T15:35:49 of course, I say it's working and now it legitimately times out 2010-11-04T15:36:42 :) 2010-11-04T15:36:56 I was just faking a feeling I've had for like a 1000 times 2010-11-04T15:37:10 yeah I've done that a few times 2010-11-04T15:37:47 i've got a great idea for a bot 2010-11-04T15:37:57 however, it is not the type of bot that plays this game 2010-11-04T15:39:39 JamesMG: if you set Console.In - make sure a) it's a static object whatever needs to can reach b) call NewStreamName.ReadLine() instead of Console.ReadLine() 2010-11-04T15:39:58 b is optional but solves a lot of potential issues 2010-11-04T15:41:11 Yeah I think I'm mangling scope of this thing somewhere 2010-11-04T15:41:23 i've got brilliant ideas for my bot 2010-11-04T15:41:29 but i actuallly have to get down to make it work 2010-11-04T15:41:31 :/ 2010-11-04T15:42:10 or misusing BufferedStream, since it seems to like timing out after 10-15 turns 2010-11-04T15:43:57 you have 2 steps to wrap it 2010-11-04T15:44:07 if you just set Console.In to something then it waits for you to write to it 2010-11-04T15:44:30 like my tcp injector sets console.in to a stringwriter and pumps to it then i can just console.readline in the main class w/o caring 2010-11-04T15:44:47 if you want to wrap the whole thing you have to set the input to the stream also 2010-11-04T15:45:00 eg BufferedStream blah = Console.GetInput() 2010-11-04T15:45:05 then Console.In = blah 2010-11-04T15:45:09 check syntax on google but that's essentially it 2010-11-04T15:46:40 you could try this also - .NET 2.0 compliant, sets height and width for buffer (SetBufferSize might work also): http://msdn.microsoft.com/en-us/library/system.console.setbuffersize%28v=VS.80%29.aspx 2010-11-04T15:46:49 I've taken the BufferedStream out, it's just using a StreamReader now 2010-11-04T15:47:14 new StreamReader(Console.GetInputStream()) and then Console.SetIn(that) 2010-11-04T15:48:27 seems to accept the game state that crashed it before 2010-11-04T15:48:58 oh sorry yea i meant bytestream not bufferedstream 2010-11-04T15:49:03 iirc bufferedstream is java lol 2010-11-04T15:49:15 streamreader works too 2010-11-04T15:49:18 C# has one too apparently, but it crashes if you use it :) 2010-11-04T15:49:34 *** iris1 has quit IRC (Ping timeout: 276 seconds) 2010-11-04T15:50:22 *** Florian95 has quit IRC (Ping timeout: 276 seconds) 2010-11-04T15:50:56 *** AJC has joined #aichallenge 2010-11-04T15:51:19 thanks for the help demerzel_ 2010-11-04T15:51:30 sure - i use it for my tcpinjector lol 2010-11-04T15:51:39 used* 2010-11-04T15:51:45 I'm about 1 more mono bug away from porting to something else :( 2010-11-04T15:52:56 here's the code for a cleaned up tcp injection btw - i really like running it directly from VS : http://pastebin.com/nQzAC4qW 2010-11-04T15:53:07 i was looking at it anyway to make sure i wasn't steering you wrong 2010-11-04T15:54:16 handy, thanks 2010-11-04T16:00:05 woah it appears that that piece of code works 2010-11-04T16:00:27 whether its correct, is a different matter 2010-11-04T16:01:30 what piece of code 2010-11-04T16:01:44 basically move selection 2010-11-04T16:01:59 but it doesn't have any offensive code yet 2010-11-04T16:02:00 nice 2010-11-04T16:02:05 still working out defense bugs 2010-11-04T16:02:16 at least it's tryign to defend itself just slightly off 2010-11-04T16:02:27 yea but i need some offensive code to play, otherwise it'll never try and take an enemy planet :P 2010-11-04T16:02:36 or a neutral 2010-11-04T16:06:39 *** bhasker_ has joined #aichallenge 2010-11-04T16:07:28 *** bhasker__ has joined #aichallenge 2010-11-04T16:07:55 *** bhasker has quit IRC (Ping timeout: 276 seconds) 2010-11-04T16:09:13 *** Utkarsh has quit IRC (Ping timeout: 276 seconds) 2010-11-04T16:10:42 *** bhasker_ has quit IRC (Ping timeout: 240 seconds) 2010-11-04T16:14:21 *** Utkarsh has joined #aichallenge 2010-11-04T16:15:01 Top 10 players: bocsimacko(4239), george(4106), dmj111(4051), GreenTea(4039), Accoun(4019), shangas(4005), Raschi(3983), felixcoto(3978), jimrogerz(3961), Hazard(3940) 2010-11-04T16:20:10 *** boegel has quit IRC (Quit: Leaving) 2010-11-04T16:20:52 *** sigh has joined #aichallenge 2010-11-04T16:21:24 *** Accoun has quit IRC () 2010-11-04T16:24:59 bhasker__: how is it comming along 2010-11-04T16:25:12 horribly lol 2010-11-04T16:25:17 my bot is getting worse by the minute 2010-11-04T16:25:20 :-\ 2010-11-04T16:25:20 :/ 2010-11-04T16:26:35 bhasker__: it's other bots getting better! 2010-11-04T16:26:49 lol no the changes i am making are making it worse 2010-11-04T16:26:50 :-\ 2010-11-04T16:27:02 well it can be hard to predict 2010-11-04T16:27:08 how are you testing? 2010-11-04T16:27:09 well its doing dumb stuff 2010-11-04T16:27:11 :) 2010-11-04T16:27:16 which it wasn't doing earlier 2010-11-04T16:27:33 i really need to run more exhaustive tests 2010-11-04T16:27:50 yeah, its a lot of waiting for results 2010-11-04T16:28:05 currently @ 44 ranking :D 2010-11-04T16:28:08 nice 2010-11-04T16:28:21 but i do expect to get much higher though 2010-11-04T16:28:31 however it takes loooooads of time -.- 2010-11-04T16:29:22 not a very interesting game but the graph at the bottom is neat: http://72.44.46.68/canvas?game_id=856433 2010-11-04T16:29:23 *** Rubicon-|-Cross has quit IRC (Quit: Leaving) 2010-11-04T16:29:48 also my bot is obviously bugged some more 2010-11-04T16:32:08 i have a new idea 2010-11-04T16:32:20 its just that it'll take a massive amount of time to implement 2010-11-04T16:32:31 Error323: the coding takes time, or the leveling up? 2010-11-04T16:32:42 AJC: both :-) 2010-11-04T16:32:42 Error323: hehe, I know what you mean. 2010-11-04T16:33:00 I'm on my tenth "idea" that requires major changes and going back to basics debugging against ragebot 2010-11-04T16:33:26 well, maybe not literally 10, but you know what I mean. 2010-11-04T16:33:37 this time, it's the last one... 2010-11-04T16:33:37 hehe 2010-11-04T16:33:42 haha 2010-11-04T16:33:58 it gets worse once your work stats to pay off 2010-11-04T16:34:08 then you _really_ want a good revisioning system 2010-11-04T16:34:20 it's easier to put it down when you are making it worse, trust me I know 2010-11-04T16:35:18 *** bhasker has joined #aichallenge 2010-11-04T16:35:20 *** Accoun has joined #aichallenge 2010-11-04T16:35:42 Error323: using git 2010-11-04T16:35:51 AJC: excellent choice ;) 2010-11-04T16:35:54 so am i 2010-11-04T16:35:59 Error323: I don't like git. 2010-11-04T16:36:02 it's obtuse. 2010-11-04T16:36:05 you are crazy 2010-11-04T16:36:16 <- hg (mercurial) 2010-11-04T16:36:24 like doing stupid shit like paginating diffs by default 2010-11-04T16:36:25 WTF? 2010-11-04T16:36:42 it is the most awesome system for non-binary data in the world 2010-11-04T16:36:52 paginating? 2010-11-04T16:36:55 demerzel_: ping 2010-11-04T16:36:55 git is awesome! 2010-11-04T16:37:00 Error323: have you used bzr, hg? 2010-11-04T16:37:17 vim vs emacs 2010-11-04T16:37:18 i've used both, i prefer git hands-down 2010-11-04T16:37:19 * Zannick ducks 2010-11-04T16:37:22 bzr? linky? 2010-11-04T16:37:22 :D 2010-11-04T16:37:30 amstan: ? 2010-11-04T16:37:37 Error323: git diff MyBot.py requires you to press space to continue. 2010-11-04T16:37:42 http://whygitisbetterthanx.com/ 2010-11-04T16:37:45 demerzel_: i was thinking of contacting the guy who made your adblock list, make him fix it 2010-11-04T16:37:52 AJC: no it doesn't 2010-11-04T16:37:55 AJC: pipe it into a different pager 2010-11-04T16:38:00 j&k 2010-11-04T16:38:06 it does this by default. 2010-11-04T16:38:09 *** bhasker__ has quit IRC (Ping timeout: 272 seconds) 2010-11-04T16:38:17 it depends what pager you have setup as default 2010-11-04T16:38:18 it has vim bindings for scrolling 2010-11-04T16:38:20 amstan: good luck with that - it's a pretty standard list, it should qualify it with the url but i have no idea how you'd go about that 2010-11-04T16:38:21 haha svn, perforce 2010-11-04T16:38:22 raedwulf: I don't want to page it, it's a non-unix like default behavior. 2010-11-04T16:38:34 demerzel_: can you tell me which line was offensive? 2010-11-04T16:38:41 I can type "| less" myself thanks. 2010-11-04T16:38:49 amstan: you mean which line in http or which line in the list? 2010-11-04T16:38:53 hehe 2010-11-04T16:38:57 demerzel_: the list 2010-11-04T16:39:00 yea one sec 2010-11-04T16:39:02 demerzel_: https://secure.fanboy.co.nz/fanboy-adblock.txt 2010-11-04T16:39:07 that one? 2010-11-04T16:39:17 well... as a terminal user i just open a second on my other screen 2010-11-04T16:39:29 demerzel_: because i can't find "##.leaderboard" in it, only "##.leaderboard2" 2010-11-04T16:39:34 and when i quit the diff, my normal terminal is back. imo its better 2010-11-04T16:39:47 but i guess thats differs per person :P 2010-11-04T16:39:56 yea i quite like vimpager :P 2010-11-04T16:40:24 Zannick: yes its becoming more like vim vs non-vim atm tho 2010-11-04T16:40:26 lol 2010-11-04T16:40:48 i mean 2010-11-04T16:40:54 *I* would rather use vim than svn 2010-11-04T16:40:56 i just love the branching, cherry-picking and merging with blazing speed and very good algorithms 2010-11-04T16:41:05 Zannick: i agree :P 2010-11-04T16:41:07 amstan: 10173 2010-11-04T16:41:09 qualitatively 2010-11-04T16:41:09 :P 2010-11-04T16:42:28 amstan: you could just change the name to ranklist or ai-leaderboard that woudl be a much faster fix 2010-11-04T16:42:32 classname that is 2010-11-04T16:42:43 demerzel_: but then i have to get the svn and do all that 2010-11-04T16:42:50 ugh... this is easier.. lol 2010-11-04T16:43:02 Error323: git is a bit retarded at merging though, so far in my experience I had better luck with SVN 1.6 even. it's not been horrible, just not as good as fanboi's claim :-) 2010-11-04T16:43:18 :o 2010-11-04T16:43:21 o_O 2010-11-04T16:43:21 o_O 2010-11-04T16:43:48 hg merge is nice 2010-11-04T16:43:52 AJC: ugh.. you're one of those guys... that hates |less automatically 2010-11-04T16:43:54 Error323: but IMHO it's the commandline and feedback that are worst, it's simply not designed with users in mind. 2010-11-04T16:44:08 hmmm 2010-11-04T16:44:32 amstan: unix has a philosophy of combining tools together to create complex stuff, why go against that? no other program does it, it's counter intuitive. 2010-11-04T16:44:41 AJC: man 2010-11-04T16:44:58 :o 2010-11-04T16:45:32 AJC: http://www.newartisans.com/blog_assets/git.from.bottom.up.pdf I read this, and then I knew why it was so awesome, it really helps with your understanding of howto use it. Read it when you have time :) 2010-11-04T16:45:37 i'm pretty sure the line of thought was ... most people want a pager for git log, so let's add it 2010-11-04T16:45:55 man always struck me as weird 2010-11-04T16:46:05 amstan: hehe, like there are not enough pager implementations in the world. 2010-11-04T16:46:17 you can select the pager implementation... 2010-11-04T16:46:18 AJC: it's just using normal less, they didn't make their own 2010-11-04T16:46:18 but it makes sense in the days before scroll bar widgets 2010-11-04T16:46:19 i created an alias for hg log to use |less 2010-11-04T16:46:20 :/ 2010-11-04T16:46:30 AJC: it's in the config files for git, you can change it to use cat instead 2010-11-04T16:46:44 export PAGER=cat 2010-11-04T16:46:50 or something should do the trick as well 2010-11-04T16:46:57 i can't remember the env variable tho 2010-11-04T16:47:14 http://paste.pocoo.org/show/286205/ 2010-11-04T16:47:18 yes its PAGER 2010-11-04T16:47:23 mega1, can I ask if your bot is deterministic or not? 2010-11-04T16:47:47 you just kinda did :P 2010-11-04T16:48:16 no I asked if I could ask. If he says I can't ask I won't ask :P 2010-11-04T16:48:26 there was a kernel mailing list discussion about this 2010-11-04T16:48:57 what you really want to know is will he answer, I don't think there's much question as to whether you're capable of forming the question 2010-11-04T16:49:34 there we go: http://kerneltrap.org/mailarchive/git/2008/12/17/4443664 2010-11-04T16:50:15 Migi32: read the channel topic. "Don't ask to ask." :P 2010-11-04T16:50:36 it does? 2010-11-04T16:50:42 oh lol that's what that means 2010-11-04T16:50:50 oh no I'm sorry! 2010-11-04T16:50:57 !roulette 2010-11-04T16:51:02 it also says "No stupid questions" which is hilariously ambiguous 2010-11-04T16:51:03 @roulette 2010-11-04T16:51:03 Migi32: *click* 2010-11-04T16:51:11 Migi32: lol, don't worry, it's for stupid people that come in here and "may i ask a question" 2010-11-04T16:51:13 @roulette 2010-11-04T16:51:13 Migi32: *click* 2010-11-04T16:51:14 @roulette 2010-11-04T16:51:14 Migi32: *click* 2010-11-04T16:51:16 @roulette 2010-11-04T16:51:16 Migi32: *click* 2010-11-04T16:51:17 @roulette 2010-11-04T16:51:17 *** Migi32 was kicked by contestbot (BANG!) 2010-11-04T16:51:18 * contestbot reloads and spins the chambers. 2010-11-04T16:51:22 *** Migi32 has joined #aichallenge 2010-11-04T16:51:22 ? 2010-11-04T16:51:23 @roulette 2010-11-04T16:51:23 Zannick: *click* 2010-11-04T16:51:26 or.. you can just go suicide 2010-11-04T16:51:27 lol 2010-11-04T16:51:30 ... 2010-11-04T16:51:46 Wow, Migi32, you have some good luck 2010-11-04T16:51:55 @roulette 2010-11-04T16:51:55 raedwulf: *click* 2010-11-04T16:52:03 k.. stop playing with that 2010-11-04T16:52:14 why is stuff like that even in there 2010-11-04T16:52:15 yeah. I had like 5 *click*'s already so I was feeling lucky :P 2010-11-04T16:52:16 -.- 2010-11-04T16:52:27 <|:>{o> 2010-11-04T16:52:31 you're right.. 2010-11-04T16:52:35 @botdoc 2010-11-04T16:52:36 amstan: botdoc = (#1) http://ubottu.com/stdin/supydocs/plugins/, or (#2) http://supybook.fealdia.org/devel/, or (#3) http://sourceforge.net/apps/mediawiki/gribble/index.php?title=Supybot_Resources 2010-11-04T16:52:47 haha I tihnk it's hilarious 2010-11-04T16:52:48 @apropos roulette 2010-11-04T16:52:48 amstan: Games roulette 2010-11-04T16:52:54 should i disable it? 2010-11-04T16:52:59 no 2010-11-04T16:53:00 nevah! 2010-11-04T16:53:01 @roulette 2010-11-04T16:53:01 *** rabidus was kicked by contestbot (BANG!) 2010-11-04T16:53:02 * contestbot reloads and spins the chambers. 2010-11-04T16:53:02 yes 2010-11-04T16:53:04 hahaha 2010-11-04T16:53:06 *** rabidus has joined #aichallenge 2010-11-04T16:53:08 wtf! 2010-11-04T16:53:09 go for it 2010-11-04T16:53:14 first time lucky :P 2010-11-04T16:53:19 nah, it's hilarious, nvm 2010-11-04T16:53:23 I say no 2010-11-04T16:53:25 make it ban people :P 2010-11-04T16:53:32 i should 2010-11-04T16:53:33 lol 2010-11-04T16:53:35 xD 2010-11-04T16:53:37 kickban 2010-11-04T16:53:40 yea, make it more interesting :P 2010-11-04T16:54:33 but anyway back on topic, mega1 is your bot deterministic? 2010-11-04T16:55:05 i think he's afk 2010-11-04T16:55:13 you can always do @tell later 2010-11-04T16:55:24 *** sigh has quit IRC (Remote host closed the connection) 2010-11-04T16:55:54 @tell rabidus hi, i love you 2010-11-04T16:56:00 :( 2010-11-04T16:56:28 oh, it works 2010-11-04T16:56:44 it's ok I'll just ask him again when he comes back online 2010-11-04T16:57:01 @tell rabidus you are the best because you wrote something @ aichallenge 2010-11-04T16:57:07 sigh :( 2010-11-04T16:57:49 rabidus are you cheating on your bot with contestbot? 2010-11-04T16:57:52 btw, how does @tell later work? It tells the person something when he comes online, but if he's already online but afk it doesn't work, right? 2010-11-04T16:58:16 I mean, it will work but you could just as well tell the person directly? 2010-11-04T16:58:41 i think it waits until it sees him become active 2010-11-04T17:00:08 there's thing call'd idletime 2010-11-04T17:01:29 *** Blkt has quit IRC (Read error: Connection reset by peer) 2010-11-04T17:03:43 *** hellman has quit IRC (Read error: Connection reset by peer) 2010-11-04T17:04:10 *** Itkovian has joined #aichallenge 2010-11-04T17:04:51 @ranking 2010-11-04T17:04:52 JamesMG: Top 10 players: bocsimacko(4227), george(4086), GreenTea(4042), dmj111(4038), shangas(4010), Accoun(3982), Raschi(3980), felixcoto(3967), jimrogerz(3956), Hazard(3953) 2010-11-04T17:05:02 soon i'll be in there :P 2010-11-04T17:05:11 Migi32: mostly deterministic 2010-11-04T17:05:45 mostly? So there is some randomness? 2010-11-04T17:06:10 I've been watching some of your games and some moves I really couldn't find a reason for other than a byproduct of randomness 2010-11-04T17:08:02 Migi32: when they say something in the channel, the bot sends you all the messages that were sent, in a pm 2010-11-04T17:08:19 rabidus: did you see this? 2010-11-04T17:08:21 @predict 2010-11-04T17:08:21 I predict that j3camero will win! 2010-11-04T17:08:26 oh yeah! 2010-11-04T17:08:39 :o 2010-11-04T17:08:40 lol 2010-11-04T17:08:43 lol, that's not even possible :P 2010-11-04T17:08:44 @predict 2010-11-04T17:08:44 I predict that eburnette will win! 2010-11-04T17:08:50 there.. that is 2010-11-04T17:08:50 it's never me.. 2010-11-04T17:09:05 I predict that contestbot will win! 2010-11-04T17:09:16 amstan: is that random? or based on elo 2010-11-04T17:09:21 or sth 2010-11-04T17:09:23 random people in the chan 2010-11-04T17:09:31 lol 2010-11-04T17:09:32 i see 2010-11-04T17:09:57 @predict 2010-11-04T17:09:57 I predict that dmj111 will win! 2010-11-04T17:10:03 damn, so true 2010-11-04T17:10:30 well unless we do sth briljant i predict bocsimacko will win 2010-11-04T17:10:31 @eightball what are my chances of winning? 2010-11-04T17:10:31 lavalamp_: No chance. 2010-11-04T17:10:36 :( 2010-11-04T17:10:37 BACK TO WORK =) 2010-11-04T17:10:44 lol 2010-11-04T17:11:05 haha, harsh 2010-11-04T17:11:13 @eightball how about making top ten? 2010-11-04T17:11:13 lavalamp_: You're kidding, right? 2010-11-04T17:11:18 doh 2010-11-04T17:11:23 burn! 2010-11-04T17:11:26 contestbot is hateful today 2010-11-04T17:11:27 JamesMG: An error occured while trying to show the previous error. 2010-11-04T17:11:51 and it was right about golang getting installed, too... :/ 2010-11-04T17:12:41 @eightball how about top eleven? 2010-11-04T17:12:41 Zannick: It is possible. 2010-11-04T17:12:45 woohoo! 2010-11-04T17:12:46 aha 2010-11-04T17:12:47 >.> 2010-11-04T17:13:38 *** Mekanik has quit IRC (Quit: WeeChat 0.3.3) 2010-11-04T17:15:01 Top 10 players: bocsimacko(4229), george(4088), GreenTea(4044), dmj111(4041), shangas(4013), Accoun(3985), Raschi(3983), felixcoto(3969), jimrogerz(3958), Hazard(3955) 2010-11-04T17:16:49 *** Appleman1234 has joined #aichallenge 2010-11-04T17:18:14 *** fawek has joined #aichallenge 2010-11-04T17:19:41 @top11 2010-11-04T17:19:42 Frontier: I do not know about 'top11', but I do know about these similar topics: 'tcp' 2010-11-04T17:19:54 @commands 2010-11-04T17:19:54 Frontier: action, add, alert, alias, allyourspace, announce, announce add, announce list, announce remove, any, aol, apply, apropos, at, author, azn, ban add, ban list, ban remove, base, binary, bold, bye, cache, calc, call, capabilities, capability add, capability list, capability remove, capability set, capability setdefault, capability unset, capitalize, change, changename, channel, channels, (5 more messages) 2010-11-04T17:20:15 @more 2010-11-04T17:20:15 Frontier: channelstats, chr, cmd, coin, color, colorize, command, commands, concat, config, connect, contributors, convert, countargs, cpu, crypt, ctime, cut, cycle, decode, default, defaultcapability, defaultplugin, dehalfop, deop, devoice, dice, dict, dictionaries, disable, disconnect, dns, do, doctype, driver, echo, eightball, elapsed, enable, encode, errno, export, fetch, field, fight, fit, (4 more messages) 2010-11-04T17:20:21 @more 2010-11-04T17:20:21 Frontier: flush, forget, format, fortune, game, get, ghost, gnu, google, grab, halfop, headers, hebrew, help, hexip, hexlify, hi, hostmask, hostmask add, hostmask list, hostmask remove, icalc, identify, ignore, ignore add, ignore list, ignore remove, info, insert, invite, ircquote, jeffk, join, kban, key, kick, last, latency, learn, leet, len, levenshtein, limit, list, lithp, ln, load, lobotomy add, (3 more messages) 2010-11-04T17:20:24 Frontier: do thin in a PM 2010-11-04T17:20:26 pls 2010-11-04T17:20:30 this* 2010-11-04T17:20:52 PM chat with contestbot? 2010-11-04T17:20:56 yes 2010-11-04T17:21:08 good idea, didnt even think of that 2010-11-04T17:24:16 @google black small dick 2010-11-04T17:24:17 rabidus: PinkyDicks.com - Pinky Dicks - The Worlds #1 Small Dick Site: ; Son has unusually small penis, worried --Doctors Lounge(TM): ; The Great Black Dick Hoax - Encyclopedia Dramatica: ; Penis Information - Penis Size Myths And Facts: (2 more messages) 2010-11-04T17:24:39 um 2010-11-04T17:24:45 well 2010-11-04T17:24:46 wut 2010-11-04T17:24:47 :D 2010-11-04T17:25:54 @google troll irc 2010-11-04T17:25:55 Frontier: BBC - h2g2 - Flaming and Trolling: ; How to Deal with an Internet Troll - wikiHow: ; Top 10 Trolls in Internet History - Online Best Latest Price ...: ; Tor Project: Abuse FAQ: hmm 2010-11-04T17:29:36 contestbot: whoami 2010-11-04T17:29:36 Zannick: I don't recognize you. 2010-11-04T17:29:44 @help whoami 2010-11-04T17:29:44 Zannick: (whoami takes no arguments) -- Returns the name of the user calling the command. 2010-11-04T17:30:01 well, it doesn't exactly return my name :P 2010-11-04T17:30:21 *** Naktibalda has joined #aichallenge 2010-11-04T17:32:58 Migi32: it's only source of randomness is sorting of equal values 2010-11-04T17:33:05 *** ChanServ sets mode: +o Zannick 2010-11-04T17:33:14 @whoami 2010-11-04T17:33:14 Zannick: I don't recognize you. 2010-11-04T17:33:17 ah ok. Strange. 2010-11-04T17:33:44 *** Zannick sets mode: -o Zannick 2010-11-04T17:33:46 * Zannick shrug 2010-11-04T17:33:58 guess i can't turn that off from here 2010-11-04T17:34:24 so all I can say now is that I have no idea what floats your bot :P 2010-11-04T17:34:31 *** hightime has quit IRC (Remote host closed the connection) 2010-11-04T17:35:34 bhasker: ur bot is now scaring me... 2010-11-04T17:36:31 Zannick: you have to login for the bot in order to work 2010-11-04T17:36:35 Zannick: only i have an account 2010-11-04T17:36:43 it's used for stuff like settings and so on 2010-11-04T17:36:44 amstan: yeah, i just figured that out 2010-11-04T17:39:25 *** Azrathud has joined #aichallenge 2010-11-04T17:41:20 How do you start a pm session? 2010-11-04T17:41:24 <- IRC noob 2010-11-04T17:41:33 /msg nick 2010-11-04T17:41:39 thx 2010-11-04T17:41:46 /msg nick message 2010-11-04T17:41:48 :) 2010-11-04T17:42:15 hm didn't open a new window 2010-11-04T17:42:21 in what client 2010-11-04T17:42:31 ah no it usually doesn't. Only when people respond. 2010-11-04T17:42:32 the web one 2010-11-04T17:42:52 hm well contestbot responded... 2010-11-04T17:42:57 not in a new window? 2010-11-04T17:43:17 nope 2010-11-04T17:43:40 i don't really know how the web one works 2010-11-04T17:43:45 strange, here neither, on XChat 2010-11-04T17:43:48 oh well 2010-11-04T17:44:04 but that might have done a notice instead of a query 2010-11-04T17:44:11 * Zannick shrug 2010-11-04T17:44:22 yeah, changed all personal replies to queries 2010-11-04T17:44:53 oh yeah, i didn't even notice he was not using queries in my pm session 2010-11-04T17:46:35 some clients won't open a new window on /msg, but wait until they receive one back 2010-11-04T17:46:58 so if a notice is used instead, it is placed in the channel window 2010-11-04T17:47:06 that's what i expect you saw in the web client 2010-11-04T17:47:53 yeah I guess so 2010-11-04T17:48:11 think I need to tell my bot to stop sniping planets clear across the map again 2010-11-04T17:48:54 I may have made my bot finally understand that some goals are worse than doing nothing, but it's too buggy to tell atm 2010-11-04T17:49:28 I need to merge all these state changes that I've made into my treebot 2010-11-04T17:49:34 Error323 fixed a bug:) 2010-11-04T17:49:40 maybe it'll win more than 10% of its games then 2010-11-04T17:51:17 *** rebelxt has left #aichallenge 2010-11-04T17:51:37 *** Sarin has quit IRC (Remote host closed the connection) 2010-11-04T17:51:46 *** phreeza has quit IRC (Ping timeout: 245 seconds) 2010-11-04T17:51:59 I almost got it working this morning but my defense code has a major bug still 2010-11-04T17:54:00 (I paused work on my minimax bot yesterday and started a simple iterative one to get the rest of the bugs out of my framework) 2010-11-04T17:54:22 ah 2010-11-04T17:54:43 the minimax makes it sooooo hard to debug anything 2010-11-04T17:54:49 yeah it does 2010-11-04T17:54:52 bhasker: no shit ;) 2010-11-04T17:55:08 stop that ^_^ 2010-11-04T17:55:12 *** sigh has joined #aichallenge 2010-11-04T17:55:20 ufh i'm in the middle of a massive refactor 2010-11-04T17:55:20 going to run another version now with one bug fix lets see how it does 2010-11-04T17:55:38 actually the funny part is the previous bug fix makes no sense to me 2010-11-04T17:55:41 but it seems to work 2010-11-04T17:55:42 :-\ 2010-11-04T17:56:11 holy crap man, it keeps on climbing 2010-11-04T17:56:24 :d 2010-11-04T17:56:35 0.19 starting up now 2010-11-04T17:56:57 http://tindeck.com/listen/ubme kids with keyboards, obama, drugs, dead aminals, lawyers, amazing dialog. 2010-11-04T17:58:02 first game against JamesMG 2010-11-04T17:58:39 which of mine? 2010-11-04T17:59:33 JamesMG.dev 2010-11-04T17:59:40 3rd game lost to shampoo :-\ 2010-11-04T17:59:43 stupid shampoo 2010-11-04T17:59:44 !@#!@# 2010-11-04T17:59:54 hhahahaha 2010-11-04T17:59:57 yeeep 2010-11-04T18:00:02 i know that feeling 2010-11-04T18:00:06 all too well -,- 2010-11-04T18:00:19 so that's how it feels to get owned by a guy named shampoo 2010-11-04T18:00:26 is it allowed to save information about the 100 maps in the bot? 2010-11-04T18:00:38 Frontier: sure 2010-11-04T18:00:54 I mean... could I just store all 100 optimal starting moves and wont get banned? 2010-11-04T18:01:21 Frontier: that won't be effective though 2010-11-04T18:01:24 it's legitimate 2010-11-04T18:01:29 Frontier: the maps might change without notice 2010-11-04T18:01:35 but not useful if new maps are used 2010-11-04T18:01:37 *** AJC has left #aichallenge 2010-11-04T18:01:48 Frontier: you can do it.. but it's not the point 2010-11-04T18:01:56 *** rebelxt has joined #aichallenge 2010-11-04T18:02:02 it's the same thing as making kids in school memorize stuff 2010-11-04T18:02:07 instead of actually teaching them to think 2010-11-04T18:02:08 amstan: the server is now running only those 100 maps? 2010-11-04T18:02:25 amstan: I have this doubt, that the maps will change all that much, but if there are new maps, I could still checksum them and add the information to the next upload 2010-11-04T18:02:49 Frontier: what if we get new maps after submissions close? 2010-11-04T18:02:53 we did that last conteset 2010-11-04T18:02:55 ah, but: what if you can't upload at that point? 2010-11-04T18:03:04 I think we know that the finals will be run with different maps after submissions are over 2010-11-04T18:03:22 so like.. it could be same sized maps, but different seeds 2010-11-04T18:03:56 amstan: will the intermediate results be shown? 2010-11-04T18:03:58 amstan: if you guys are going to change the number of planets you really need to announce that... I'm not the only one depending on it... 2010-11-04T18:04:15 amstan: I have analyzed quite a few starting moves and it looks like I could get a slight advantage out of these precalculations. Remember these maps where the starting planets are really distant from each other? There is plenty of time for 'dump' replay of optimal moves. 2010-11-04T18:04:20 lavalamp_: number of planets is variable, isn't it? 2010-11-04T18:04:26 I'm depending on you increasing planets to 1000 so I can win 2010-11-04T18:04:29 amstan: always 23 2010-11-04T18:04:33 oh.. 2010-11-04T18:04:36 Zannick: true..., but still 70% of the maps would be known to my bo 2010-11-04T18:04:39 server has more than 100 maps 2010-11-04T18:04:39 that doesn't sound like fun 2010-11-04T18:04:41 my code says if Planets.count > 999 return; 2010-11-04T18:04:46 man tcp is a tough proving ground 2010-11-04T18:05:03 iot's god tho 2010-11-04T18:05:07 *** nigg_ has joined #aichallenge 2010-11-04T18:05:10 yeesh - it's good tho 2010-11-04T18:05:18 tcp server has 999 but a handful of those are bad 2010-11-04T18:05:46 what counts as bad? high likelihood of draw? 2010-11-04T18:05:57 sigh: yeah, or rps maps 2010-11-04T18:06:15 i /think/ i am bug-free finally 2010-11-04T18:06:17 rps maps aren't bad! 2010-11-04T18:06:20 now i just get my ass handed to me 2010-11-04T18:06:20 nvm 2010-11-04T18:06:22 you're right 2010-11-04T18:06:23 sorry if i missed this.. will there be a new set of maps generated for the final run? 2010-11-04T18:06:23 sorry 2010-11-04T18:06:24 amstan: You won't change the game rules in this last month. At least I cant imagine that. Some bots really have code that can only run efficiently in under 1 second if there are 23 planets at most. 2010-11-04T18:06:28 there are proven rps maps? 2010-11-04T18:06:40 yep 2010-11-04T18:07:11 *** Bobng has quit IRC (Ping timeout: 245 seconds) 2010-11-04T18:07:20 sigh: i mean 'bad' as in fails the too_close condition the generator has 2010-11-04T18:07:32 Zannick: did you answer my question? sorry, not sure ;-) 2010-11-04T18:07:48 those maps were made before that condition was applied to the starting planets and the center 2010-11-04T18:07:49 Zannick: ah, i see 2010-11-04T18:08:05 sigh: mcleopold had a very good rps map 2010-11-04T18:08:20 rebelxt: no, i didn't, but i believe amstan said yes earlier 2010-11-04T18:08:43 Zannick: ok, got it 2010-11-04T18:08:57 how do you advanced game theory guys handle the probability stuff? just take the highest one? 2010-11-04T18:09:19 @google "mixed strategies" 2010-11-04T18:09:20 lavalamp_: Strategy (game theory) - Wikipedia, the free encyclopedia: ; Mixed Strategies: ; Mixed Strategies: ; Mixed Strategy in Game Theory - Game Theory .net: (2 more messages) 2010-11-04T18:09:22 the rps map: http://pastebin.com/ZZK3TkzF 2010-11-04T18:09:54 amstan: At least it would be fair if we knew by now what that the game rules and planet distances/numbers/symmetry will stay the same. People who try to support any number of planets or asymetric maps will have a disatvantage, because they miss optimization opportunities. 2010-11-04T18:10:04 *** JamesGecko has joined #aichallenge 2010-11-04T18:10:22 what is most likely to happen is the map generator will simply be run some more 2010-11-04T18:10:31 yea i know what a mixed strategy is 2010-11-04T18:10:32 which will generate maps with the same conditions 2010-11-04T18:10:34 I'm counting on that 2010-11-04T18:10:42 i am asking if you end up with 2 actions prob 75/25 do you always take 75 2010-11-04T18:10:49 demerzel_: choose randomly 2010-11-04T18:10:51 Frontier: i told you, probably the seeds will change 2010-11-04T18:10:51 Zannick: I think most of us silently assume that ^^ 2010-11-04T18:11:18 75% of the time do the first one. 2010-11-04T18:11:32 Frontier: i think that answers your statement above, then 2010-11-04T18:11:33 amstan: So the map with ID 1 will have a different layout or do you mean, you create more maps 200...900? 2010-11-04T18:11:49 Frontier: the map number doesn't matter, you're not told it 2010-11-04T18:11:55 *** Appleman1234 has quit IRC (Read error: Connection reset by peer) 2010-11-04T18:11:56 I think amstan is hiding now :) 2010-11-04T18:12:04 we could name the maps if we liked 2010-11-04T18:12:16 map_general_custer.txt 2010-11-04T18:12:21 I can think of some names for map 11... 2010-11-04T18:12:30 map_jamesmg_will_lose_on_this_map.txt 2010-11-04T18:12:38 just pick one to name that, will be accurate eventually 2010-11-04T18:12:49 Zannick: Ok, so will the seed change result in all my cached map data becoming invalid? 2010-11-04T18:12:50 hmm 2010-11-04T18:12:59 0.19 is doing quite well except for that single loss to shampoo 2010-11-04T18:13:10 Frontier: what are you caching on? 2010-11-04T18:13:22 if you're doing a straight hash of all the planet data, then i doubt it 2010-11-04T18:14:10 if are caching by id, and doing some sort of check, then also no 2010-11-04T18:14:31 because it doesn't matter how the server numbers planets, if you keep your own index 2010-11-04T18:14:51 Zannick: I am caching on initial moves, by hash over the map data. 2010-11-04T18:14:53 *** Appleman1234 has joined #aichallenge 2010-11-04T18:15:02 Top 10 players: bocsimacko(4231), george(4084), dmj111(4050), shangas(4030), jimrogerz(4005), Accoun(3981), Hazard(3978), Raschi(3961), felixcoto(3958), Xrillo(3942) 2010-11-04T18:15:05 I imagine if they generate new maps with a different seed they will be completely different maps 2010-11-04T18:15:11 so yes, hashing the map data won't change 2010-11-04T18:15:24 if you get a different map you simply won't find it in your cache 2010-11-04T18:15:38 unless you get really unlucky and there's a hash collision 2010-11-04T18:15:43 ...if not then something is wrong with the rng 2010-11-04T18:15:51 ...and what I'm trying to tell is how useful that caching would be if there are new maps after the deadline. 2010-11-04T18:16:01 *** JamesMG has quit IRC (Quit: Leaving.) 2010-11-04T18:16:10 Frontier: my guess: not very :) 2010-11-04T18:16:13 Will all maps be replaced? If yes, I can dump that idea. 2010-11-04T18:16:16 if there are only new maps, not useful 2010-11-04T18:16:22 if both, partially useful 2010-11-04T18:16:31 Zannick: thx for stating the obvious ^^ 2010-11-04T18:16:42 you won't be able to update the cache afterwards... 2010-11-04T18:17:21 you're welcome ^^, 2010-11-04T18:17:27 but my code sometimes takes 30 minutes and 2GB of ram to calculate the initial moves :( 2010-11-04T18:17:29 i honestly don't think it would be very useful 2010-11-04T18:18:02 o.o 2010-11-04T18:18:54 *** narnach has quit IRC (Remote host closed the connection) 2010-11-04T18:19:01 22*21*20... different source planet combinations for each new target planet... all that stuff that ends in an explosion of branches 2010-11-04T18:19:17 *** narnach has joined #aichallenge 2010-11-04T18:19:25 oh, you're doing everything? no wonder :P 2010-11-04T18:19:59 i'm not saying you shouldn't do this cache, i'm only saying it won't help you for maps you don't recognize 2010-11-04T18:20:02 shampoo is my bane 2010-11-04T18:22:11 *** Appleman1234 has quit IRC (Read error: Connection reset by peer) 2010-11-04T18:24:00 *** lavalamp_ has quit IRC (Quit: Page closed) 2010-11-04T18:25:40 ohh man my plan better work, otherwise i'll go mental 2010-11-04T18:32:02 *** Appleman1234 has joined #aichallenge 2010-11-04T18:36:57 *** RainCT has quit IRC (Remote host closed the connection) 2010-11-04T18:38:23 this is demoralizing http://72.44.46.68/getplayer?player=bocsimacko-b0.47 2010-11-04T18:38:26 :-\ 2010-11-04T18:38:32 *** Appleman1234 has quit IRC (Read error: Connection reset by peer) 2010-11-04T18:39:57 yeah mega1's bot forgot how to lose games ;) 2010-11-04T18:40:40 bug fixes are bad 2010-11-04T18:40:48 my bot is doing worse after bug fix:-\ 2010-11-04T18:41:16 bhasker: don't worry, same here ;-) i'm done fixing bugs 2010-11-04T18:42:16 *** demerzel_ has quit IRC (Quit: Page closed) 2010-11-04T18:45:09 well at least i can still beat ragebot 2010-11-04T18:47:23 damnit i hate it when my brain stops working >_< 2010-11-04T18:47:39 i'm not doing any more programming till tuesday 2010-11-04T18:48:00 then i'm doing non stop programming to the end 2010-11-04T18:48:11 i whish i had that time 2010-11-04T18:48:22 summer holidays rule 2010-11-04T18:48:35 i basically have till the end of february to fill 2010-11-04T18:48:53 lol 2010-11-04T18:48:59 going into winter here 2010-11-04T18:49:33 it'd suck for it to be winter over christmas and new years :( 2010-11-04T18:49:50 i always go to a big 3 day music festival over new years, that would be unbearable in the winter 2010-11-04T18:54:36 *** Appleman1234 has joined #aichallenge 2010-11-04T18:55:43 *** iris1 has joined #aichallenge 2010-11-04T18:56:47 *** rebelxt has left #aichallenge 2010-11-04T18:58:50 *** Appleman1234 has quit IRC (Read error: Connection reset by peer) 2010-11-04T19:08:13 *** other_donkey has joined #aichallenge 2010-11-04T19:08:41 antimatroid1: I don't see your bot anywhere on the rankings, what happened to it? 2010-11-04T19:08:51 *** superflit has quit IRC (Quit: superflit) 2010-11-04T19:10:29 do you wonder when the logic makes no sense but your bot is doing much better :-\ 2010-11-04T19:11:32 bhasker: yeah, I used to have that problem... now my bot doesn't get better anymore 2010-11-04T19:11:49 0.18 version of mine is ripping it up 2010-11-04T19:11:54 0.19 with the correct logic 2010-11-04T19:11:57 is in the doldrums 2010-11-04T19:11:58 :-\ 2010-11-04T19:12:15 what did you "fix"? 2010-11-04T19:12:33 my calculations for how many available ships a particular planet has 2010-11-04T19:12:37 :-\ 2010-11-04T19:12:44 *** fawek has quit IRC (Read error: Connection reset by peer) 2010-11-04T19:13:14 heh, I've had that problem too 2010-11-04T19:13:35 the logic in 0.18 makes no sense 2010-11-04T19:13:40 i can see that its probably wrong 2010-11-04T19:13:45 but its doing much better 2010-11-04T19:13:47 !! 2010-11-04T19:14:03 is it more conservative or less conservative? 2010-11-04T19:14:09 less conservative i guess 2010-11-04T19:14:18 ah yes then I understand 2010-11-04T19:14:27 the reality is i don't even know i just know its wrong 2010-11-04T19:14:28 lol 2010-11-04T19:14:28 yeah, I've found being more aggressive can help sometimes 2010-11-04T19:14:40 I had a bot which always sent all ships it had and it did quite fine actually 2010-11-04T19:14:57 0.19 is at 393 elo on tcp 2010-11-04T19:14:59 Migi32: you had ragebot too?? 2010-11-04T19:14:59 just not on maps with close starting planets 2010-11-04T19:15:02 oops 289 2010-11-04T19:15:05 0.18 is at 393 2010-11-04T19:15:07 :-\ 2010-11-04T19:15:08 Top 10 players: bocsimacko(4281), george(4078), dmj111(4073), shangas(4066), jimrogerz(4041), Accoun(4000), Hazard(3999), Xrillo(3979), felixcoto(3971), Raschi(3960) 2010-11-04T19:15:10 sigh, no it did expand as well 2010-11-04T19:15:13 almost a 100 elo points diff 2010-11-04T19:15:22 Migi32: yeah, I was joking :P 2010-11-04T19:15:43 bhasker: 393 - 393 = 0 2010-11-04T19:15:48 no no 2010-11-04T19:15:50 oh, you corrected yourself 2010-11-04T19:15:54 yes 2010-11-04T19:15:57 sorry, didn't see :) 2010-11-04T19:17:28 i guess i should let it churn a bit 2010-11-04T19:17:33 0.19 is slowly climbing up 2010-11-04T19:17:53 yeah, rankings can be very tempremental 2010-11-04T19:18:08 and depends *a lot* on who the other people in the tournament are 2010-11-04T19:18:23 true 2010-11-04T19:18:35 so you submitting a new version sigh? 2010-11-04T19:19:04 I've developed an algorithm that can identify very good attack plans really well, but the problem is that I can only evaluate about 4000 positions per second... 2010-11-04T19:19:10 I dunno... I haven't got anything significantly different 2010-11-04T19:19:18 what do you guys use to generate moves? 2010-11-04T19:19:30 (if it's any different from evaluating the moves) 2010-11-04T19:19:46 mine is a heuristic bot 2010-11-04T19:19:53 so its like take planet A 2010-11-04T19:19:56 can i acquire it 2010-11-04T19:19:58 Migi32: I just decide the best way to attack destinations independantly, then combine the plans 2010-11-04T19:19:58 launch 2010-11-04T19:20:01 rinse repeat 2010-11-04T19:20:13 thats about how complicated the logic is 2010-11-04T19:20:24 "best way" is a very crude approximation :P 2010-11-04T19:20:34 hmm... Problem is that this evaluator is kinda like a back box. Put in a position, get back a number. 2010-11-04T19:20:36 the real complication is in figuring out how many ships i can send and how many i need ot send to capture 2010-11-04T19:21:12 i swear this shampoo will be the death of me 2010-11-04T19:21:18 *** Appleman1234 has joined #aichallenge 2010-11-04T19:22:57 flag how the hell does your bot manage to beat mine so many times but always stay at 200 elo!! 2010-11-04T19:23:18 because you guys are playing other people too? 2010-11-04T19:23:34 his bot seems to beat quite a few bots a 100 elo above him :-\ 2010-11-04T19:23:56 maybe it also loses to bots below him? 2010-11-04T19:25:13 probably does 2010-11-04T19:25:17 oh 0.19 beat 0.18 2010-11-04T19:25:46 lol bhasker 2010-11-04T19:25:56 shampoo is a true mistery +) 2010-11-04T19:25:58 sigh: current source is spread across multiple folders and the server wont take that 2010-11-04T19:26:08 and i can't be effed cleaning it up when it's not even very good 2010-11-04T19:27:02 i wish the server would just let me separate my files into folders, it makes my life much better 2010-11-04T19:27:58 *** jaspervdj has quit IRC (Quit: NEVER GONNA GIVE YOU UP NEVER GONNA LET YOU DOWN) 2010-11-04T19:28:03 my bot is currently losing and winning against pretty much everyone, except always losing to bocsimacko 2010-11-04T19:28:45 and testing from piter usually gets me more often than not 2010-11-04T19:29:21 testingFromPiter is quite good 2010-11-04T19:34:46 *** demerzel_ has joined #aichallenge 2010-11-04T19:36:10 *** mceier has quit IRC (Quit: leaving) 2010-11-04T19:38:34 While I was talking about caching map information... is it only me, or does bocsimacko lose only on maps that he had not on his hard drive? the ids all seem to be above 100. 2010-11-04T19:39:28 *** rangzen has joined #aichallenge 2010-11-04T19:39:30 i dunno about that 2010-11-04T19:39:31 *** nigg_ has quit IRC (Ping timeout: 255 seconds) 2010-11-04T19:39:37 If so he will probably fail miserably when the maps are exchanged in the final round after the deadline 2010-11-04T19:39:48 sure he doesn't just very rarely lose, and there's more new maps than old maps? 2010-11-04T19:40:19 i don't think he's relying on "knowing" the map previously 2010-11-04T19:40:42 antimatroid: You can calculate some best moves for the start of the round to maximise ships or growth. Maybe something like that helps him do things in under 1 second that would require hours to be done on the fly. 2010-11-04T19:41:18 he also has to match the map to one, the bot doens't get id on starting 2010-11-04T19:41:21 that'd be costly 2010-11-04T19:41:25 i really doubt it 2010-11-04T19:42:59 match a single 10 digit double and a secondary as confirmation is necessary isnt' bad 2010-11-04T19:43:25 that's really all you'd need 2010-11-04T19:43:34 antimatroid1: a hash of the map isn't THAT costly as a lookup for a table of 100 ids 2010-11-04T19:43:36 i have all the maps in a sql db that i haven't used haha 2010-11-04T19:44:10 i did have the top 100 matches but pre-mega1 so pretty worthless now and i don't wanna clean it up 2010-11-04T19:44:58 so you could select all matches of bot A on map B for example? 2010-11-04T19:45:04 yea 2010-11-04T19:45:24 i still have the web extractor i might use it if i start feeling desperate in a couple weeks 2010-11-04T19:45:31 what i lack in creativity i can query the hell out of lol 2010-11-04T19:46:02 my initial thought was to track initial selections 2010-11-04T19:46:16 but that is insanely hard now because there are so many variables plus mixed strategies seem to be the norm 2010-11-04T19:46:38 antimatroid1: well, bocsimacko hasn't lost a single game on a map with an id <= 100. But the low ids are really rare after all so you might be right 2010-11-04T19:46:44 *** davidjliu has quit IRC (Quit: Page closed) 2010-11-04T19:46:51 *** Accoun has quit IRC () 2010-11-04T19:48:16 I guess the admin wouldn't be so happy if 3000 people loaded all the games through http-requests. 2010-11-04T19:49:13 Frontier: We need somebody to write a genetic algorithm that auto-submits bots once every hour to test different parameters. That would be ballin. 2010-11-04T19:49:14 :P 2010-11-04T19:49:34 don't really need to submit with the tcp server 2010-11-04T19:49:48 also, submitting to the main server doesn't get you as many games 2010-11-04T19:49:56 hehe. when amstan sais he might change the random seed for the maps... what does that mean? is the random seed the map id? 2010-11-04T19:49:59 ah, you were talking about something ekse 2010-11-04T19:50:15 Frontier: the seed for the random number generator 2010-11-04T19:50:29 ...and also the map id ??? 2010-11-04T19:50:33 no 2010-11-04T19:50:36 -.- 2010-11-04T19:51:12 *** Accoun has joined #aichallenge 2010-11-04T19:51:14 Frontier: are you trying to predict the random seed? :P 2010-11-04T19:51:24 the seeds for the current maps have not been saved 2010-11-04T19:51:45 the map generator doesn't even have a place to input it, though you can hack around it by importing random first and seeding from there 2010-11-04T19:52:00 ok, that was the missing information. so currently the seed is fed by system time or something and the map id just increments 2010-11-04T19:52:25 yeah, seeded by whatever the default it 2010-11-04T19:52:29 s/it/is/ 2010-11-04T19:52:56 do we know who testingFromPiter is on official server? 2010-11-04T19:53:05 i wonder how many decent bots will be put up right at the end 2010-11-04T19:53:24 mine was a top 10 worthy bot last time that we didn't get ready till like a week or two before the final tourny 2010-11-04T19:54:24 the bar has been rising pretty rapidly 2010-11-04T19:54:31 i'd be surprised if a total unknown swept in at the last second 2010-11-04T19:54:49 i wouldn't 2010-11-04T19:54:50 *** Migi32 has quit IRC (Quit: http://boinc.berkeley.edu) 2010-11-04T19:55:10 there is some incentive in having people not observe your games 2010-11-04T19:55:23 *** Appleman1234 has quit IRC (Read error: Connection reset by peer) 2010-11-04T19:55:29 yes, but you also don't get as much feedback 2010-11-04T19:55:59 trying to work in isolation is hard, I think 2010-11-04T19:56:15 it was easier for tron i'll admit that much, but i don't think it's undoable 2010-11-04T19:56:24 one can watch other peoples games 2010-11-04T19:56:30 unless they just read all the channel logs and study the top games intensively :P 2010-11-04T19:57:24 you don't think the people in this chan do that? 2010-11-04T19:57:24 i expect there's at least a few people with half decent bots that aren't openly testing them 2010-11-04T19:57:28 no 2010-11-04T19:57:38 lol i do :P 2010-11-04T19:57:39 most of the people on here are always talking about their bots on tcp 2010-11-04T19:57:52 altho i don't study top games 2010-11-04T19:57:56 there are a lot of lurkers here 2010-11-04T19:58:02 true 2010-11-04T19:58:06 i feel like i have a pretty good feel for when a strat is good by now just watching viz 2010-11-04T19:58:13 I like to suprise them every so often with 2010-11-04T19:58:15 @predict 2010-11-04T19:58:15 I predict that Cyndre will win! 2010-11-04T19:58:30 pretty much all of my ideas are available in chan logs, but i don't think it matters much about knowing ones idea for this 2010-11-04T19:58:36 the implementation is so crtitical 2010-11-04T19:58:40 ^^ 2010-11-04T19:58:48 and fun! 2010-11-04T19:59:07 i looked at this like a coding challenge but now it's an intellectual one haha 2010-11-04T19:59:09 lmao 2010-11-04T19:59:19 yeah, personally I enjoy this competition more for the exhange of ideas than anything else 2010-11-04T19:59:22 *exchange 2010-11-04T19:59:25 I havent touched it in a month 2010-11-04T19:59:29 coding challenges should be intellectual ones :P 2010-11-04T19:59:57 I'm not taking the competition too seriously now 2010-11-04T20:00:03 ok stallman 2010-11-04T20:00:12 i'm gonna put the gnu license on my source after competition 2010-11-04T20:00:14 although codejam and acm etc. are often who can match isomorphic problems the best 2010-11-04T20:00:16 every page 2010-11-04T20:00:22 *** wh1teside_ has joined #aichallenge 2010-11-04T20:00:32 which is boring 2010-11-04T20:00:41 so whats the winnning tactic? 2010-11-04T20:00:43 antimatroid1: yeah, that is annoying a bit 2010-11-04T20:01:06 but for such short competitions, you can't really do much else 2010-11-04T20:01:18 Cyndre: hack server and steal bocksimacko's code 2010-11-04T20:01:40 i think minimax will win again 2010-11-04T20:01:46 antimatroid1: project euler does a decent job of avoiding that I think 2010-11-04T20:01:56 damn, going to have to check out later - I have a Toy Story 3 to watch with kids, afk 2010-11-04T20:01:59 but their problems get a bit repetitive 2010-11-04T20:02:01 sigh: yep, i've found many neat little problems on there 2010-11-04T20:02:31 you can only prove some random property of the totient function so many times before it gets old :P 2010-11-04T20:02:38 haha yep 2010-11-04T20:02:46 anyway, i'm off to the uni to pick up old assignments 2010-11-04T20:02:50 cya 2010-11-04T20:03:15 *** wh1teside has quit IRC (Ping timeout: 240 seconds) 2010-11-04T20:03:19 what is the MIT AI challenge is that a yearly thing or what 2010-11-04T20:03:28 never heard of it but somebody on the forum mentioned it 2010-11-04T20:03:32 i've heard of topcoder but that's more just implementation trickery 2010-11-04T20:04:11 oh i forgot they do something like that, i heard about it before i would have had any idea how to do anything 2010-11-04T20:05:10 http://battlecode.mit.edu/2010/ 2010-11-04T20:06:05 why java :( 2010-11-04T20:07:08 people need a good kick in the back side sometimes 2010-11-04T20:12:11 should i celebreate the only loss on this page is to my bot http://72.44.46.68/getplayer?player=bocsimacko-b0.47 2010-11-04T20:12:12 muwahaha 2010-11-04T20:12:30 i ignore the part where he destroyed me 10 times more 2010-11-04T20:15:03 Top 10 players: bocsimacko(4265), george(4062), dmj111(4054), shangas(4038), jimrogerz(4015), Accoun(4004), Hazard(3969), felixcoto(3959), Xrillo(3939), asavis(3927) 2010-11-04T20:16:54 mit is java? 2010-11-04T20:17:06 if anybody would be strict ansi-c i would have guessed mit 2010-11-04T20:19:36 you guys get too hung up on language 2010-11-04T20:24:46 *** Accoun has quit IRC () 2010-11-04T20:26:10 *** rangzen has quit IRC (Quit: Quitte) 2010-11-04T20:30:59 *** yellowseed has joined #aichallenge 2010-11-04T20:33:00 holy shit i just turned on mega logging and realized that i've been skipping planet 0 forever 2010-11-04T20:33:14 haha wtf how did that not break anything 2010-11-04T20:33:56 :-P that's hilarious. i haven't done enough debugging with mine to uncover bugs like that, so i have no idea how many there are... 2010-11-04T20:34:40 yea it's a little disconcerting tbh lol 2010-11-04T20:35:43 i love writing in python, but i almost never take a debugger to it - i just read the code carefully, do some logging, and hope i haven't missed anything. 2010-11-04T20:35:46 *** hambone has joined #aichallenge 2010-11-04T20:35:54 hi 2010-11-04T20:35:55 can 2010-11-04T20:36:04 is anyone here 2010-11-04T20:36:05 hey hambone 2010-11-04T20:36:11 how do I IRC to somewhere 2010-11-04T20:36:14 I used the 2010-11-04T20:36:16 web client 2010-11-04T20:36:29 can i use it to irc somewhere else 2010-11-04T20:36:37 haha google mirc 2010-11-04T20:36:40 i'm sure you can 2010-11-04T20:37:08 i use it cuz it's more convenient to leave a tab open in ff than have another window open 2010-11-04T20:37:21 is there any standard 2010-11-04T20:37:22 thing 2010-11-04T20:37:35 in unix or whatever for irc 2010-11-04T20:37:37 mirc is pretty much universal 2010-11-04T20:37:44 oh idk about unix i know there is probably many 2010-11-04T20:37:48 i cannot download 2010-11-04T20:38:00 what might i expect to be installed 2010-11-04T20:38:14 idk google linux irc i'm sure you'll be overwhelmed 2010-11-04T20:38:34 probably something builtin 2010-11-04T20:38:38 even 2010-11-04T20:39:11 if you're at the command line, try "apropos irc" 2010-11-04T20:39:40 hmm 2010-11-04T20:39:41 thx 2010-11-04T20:40:13 \help 2010-11-04T20:40:17 lol 2010-11-04T20:41:05 *** sigh has quit IRC (Remote host closed the connection) 2010-11-04T20:41:19 *** sigh has joined #aichallenge 2010-11-04T20:42:03 thanks 2010-11-04T20:42:07 figured out how to connect 2010-11-04T20:42:18 *** Accoun has joined #aichallenge 2010-11-04T20:42:41 *** hambone has quit IRC (Quit: Page closed) 2010-11-04T20:42:59 *** Cold-Phoenix has quit IRC () 2010-11-04T20:44:14 *** bhasker has quit IRC (Quit: bhasker) 2010-11-04T20:49:41 *** sigh has quit IRC (Ping timeout: 245 seconds) 2010-11-04T20:51:51 *** sigh has joined #aichallenge 2010-11-04T20:58:18 *** Azrathud has quit IRC (Ping timeout: 240 seconds) 2010-11-04T20:58:40 *** amstan has quit IRC (Remote host closed the connection) 2010-11-04T20:59:24 *** amstan has joined #aichallenge 2010-11-04T20:59:24 *** ChanServ sets mode: +o amstan 2010-11-04T21:03:16 *** amstan has quit IRC (Remote host closed the connection) 2010-11-04T21:04:34 *** amstan has joined #aichallenge 2010-11-04T21:04:34 *** ChanServ sets mode: +o amstan 2010-11-04T21:05:31 *** bhasker has joined #aichallenge 2010-11-04T21:07:34 *** amstan has quit IRC (Read error: Connection reset by peer) 2010-11-04T21:08:11 *** amstan has joined #aichallenge 2010-11-04T21:08:11 *** ChanServ sets mode: +o amstan 2010-11-04T21:15:03 Top 10 players: bocsimacko(4278), george(4077), dmj111(4073), shangas(4042), jimrogerz(4005), Accoun(4001), asavis(3976), felixcoto(3961), Hazard(3959), Xrillo(3943) 2010-11-04T21:15:18 *** superflit has joined #aichallenge 2010-11-04T21:21:35 *** virdo has joined #aichallenge 2010-11-04T21:50:15 *** Appleman1234 has joined #aichallenge 2010-11-04T22:04:52 *** yellowseed has quit IRC (Ping timeout: 255 seconds) 2010-11-04T22:15:05 Top 10 players: bocsimacko(4266), george(4053), dmj111(4046), shangas(4024), jimrogerz(3981), Accoun(3981), asavis(3952), felixcoto(3937), Hazard(3928), Xrillo(3895) 2010-11-04T22:25:59 Sup sup! 2010-11-04T22:26:17 I am loving whoever made the changes to game_stats.php. Really awesome! 2010-11-04T22:26:23 2.26am 2010-11-04T22:26:26 :/ 2010-11-04T22:27:03 im sleepy 2010-11-04T22:27:07 i should go back to sleep 2010-11-04T22:27:12 Are you in England, I suppose? 2010-11-04T22:27:18 yea 2010-11-04T22:28:06 * raedwulf rolls over... laptop is beside bed >:P 2010-11-04T22:28:58 lol 2010-11-04T22:28:59 oh man 2010-11-04T22:29:49 Hmm why is it so quiet in here? 2010-11-04T22:31:39 j3camero: was a lot of talking earier 2010-11-04T22:31:57 check the logs, lots of stuff there 2010-11-04T22:32:17 Sweet 2010-11-04T22:32:26 I'm running it in screen so I can just look back. 2010-11-04T22:32:29 Same difference 2010-11-04T22:34:39 I have an analysis exam tomorrow, topics in adv. maths exam monday and pdes exam tuesday, that's why i'm quiet :P 2010-11-04T22:34:55 but then i have nothing to do but make the winning bot till the end of the contest :D 2010-11-04T22:37:14 *** Naktibalda has quit IRC (Remote host closed the connection) 2010-11-04T22:44:57 i wish there was notation to say a sequence is in a set 2010-11-04T22:45:02 like subset and element of 2010-11-04T22:45:05 sequence of 2010-11-04T22:45:27 There is in Python hehe 2010-11-04T22:45:47 Python > math 2010-11-04T22:45:52 :-P 2010-11-04T22:47:06 what notation does python use for that? 2010-11-04T22:47:13 perhaps i shall just adopt it 2010-11-04T22:53:01 python doesn't use "notation" 2010-11-04T22:53:13 it's beyond it 2010-11-04T22:53:33 i've come to love simple notation 2010-11-04T22:53:45 the more i can say with less, the better 2010-11-04T22:55:34 in time we will realize that 'strong typing' is just a leftover from our ape brains 2010-11-04T22:56:19 what's "strong typing"? 2010-11-04T22:56:29 strong typing is such an ambiguous term 2010-11-04T22:56:39 it means different things to everyone 2010-11-04T22:56:57 what are some of the common things? 2010-11-04T22:57:28 I think demerzel_ is using it as a synonym for static typing 2010-11-04T22:57:41 lol, what's that? :P 2010-11-04T22:57:58 I shall stay out of this debate. Anybody who brings up strong typing obviously has an extreme opinion on the subject. They're usually from Haskell-land, where the trolls lurk. 2010-11-04T22:58:01 static typing is where the variables have a defined type at compile tie 2010-11-04T22:58:13 i plan to learn haskell at some point 2010-11-04T22:58:15 *time 2010-11-04T22:58:33 oh yeah, i do often wish c++ wasn't like that 2010-11-04T22:59:00 c++ has shitty runtime type stuff too it's not neatly packaged like reflection but it's mashed onto c 2010-11-04T22:59:03 like most of the other stuff ;) 2010-11-04T23:00:04 wikipedia: In computer science and computer programming, a type system is said to feature strong typing when it specifies one or more restrictions on how operations involving values of different data types can be intermixed. 2010-11-04T23:00:26 i'm just messin with ya i don't care really 2010-11-04T23:00:29 so you can see how people have different "interpretations" 2010-11-04T23:00:33 but an 'easy' dev language to me is a trap 2010-11-04T23:01:01 yeah fair enough 2010-11-04T23:02:44 demerzel_: 'easy' dev language? 2010-11-04T23:02:53 python is marketed as that 2010-11-04T23:03:02 implicit typing, foreach everything 2010-11-04T23:03:41 fair enough 2010-11-04T23:05:33 foreach everything is really, really nice though... I can't see any disadvantage there :) 2010-11-04T23:06:49 what i really want is a keyboard with all the greek letters and maths symbols accesible with a key like shift 2010-11-04T23:07:08 are those characters usable in source code? 2010-11-04T23:07:15 antimatroid1: look up APL 2010-11-04T23:08:17 haha... the wikipedia page has an example of an APL keyboard: http://en.wikipedia.org/wiki/APL_(programming_language) 2010-11-04T23:09:04 it also wins the award for the most imaginative name ever 2010-11-04T23:09:06 seems kind of cool 2010-11-04T23:09:12 yeah 2010-11-04T23:09:15 *** yasith has joined #aichallenge 2010-11-04T23:09:16 but is it any good? 2010-11-04T23:09:31 haskell seems cool 2010-11-04T23:09:51 python and haskell are my biggest priorities when i can find time 2010-11-04T23:09:59 dunno, if you are interested in APL but want normal characters then try J 2010-11-04T23:10:23 nah, performance wise? 2010-11-04T23:10:33 dunno 2010-11-04T23:10:40 most of my porgramming is just doing stuff like project euler problems, or my own little things 2010-11-04T23:10:51 memory is my biggest bottleneck (probably the same for everyone) 2010-11-04T23:11:01 sigh: http://img.waffleimages.com/515239aa7c709ae221a41fff79cb70a86f40b9a0/idiocracy_interface_para_hospital.jpg 2010-11-04T23:11:02 ah, if you are looking at the euler forums you'll get used to seeing J 2010-11-04T23:11:21 demerzel_: o_O 2010-11-04T23:11:35 oops 2010-11-04T23:11:37 http://www.tru7h.org/xzzy/files/movies/idiocracy_interface_para_hospital.jpg 2010-11-04T23:11:51 APL3 keyboard 2010-11-04T23:12:04 heh 2010-11-04T23:12:25 antimatroid1: I did a vast majority of euler problems in python 2010-11-04T23:12:34 i've done all mine in c++ :P 2010-11-04T23:12:44 performance wasn't too much of an issue except for a few problems 2010-11-04T23:12:59 i don't understand why i can't just guy a keyboard with all the greek symbols on it, surely it wouldn't be that hard for some chinese manufacturer to make them and sell them 2010-11-04T23:13:07 buy** 2010-11-04T23:13:27 are they all part of unicode or whatever it is? 2010-11-04T23:13:39 (my actualy knowledge of that kind of stuff is almost non existent) 2010-11-04T23:13:47 (and my englishy abilities) 2010-11-04T23:14:00 my knowledge is enough to stay the fuck away from unicode whenever possible 2010-11-04T23:15:04 Top 10 players: bocsimacko(4241), george(4041), dmj111(4000), shangas(3996), jimrogerz(3981), Accoun(3970), asavis(3936), felixcoto(3929), Hazard(3905), Xrillo(3901) 2010-11-04T23:15:17 also, if you are running into memory bottlenecks in c++ on project euler then you need to improve your algorithms 2010-11-04T23:15:19 sigh: You should work with Java, where Unicode is the base character representation. 2010-11-04T23:15:42 *** gMonkey has joined #aichallenge 2010-11-04T23:15:48 if i ever program in java again it'll be too soon 2010-11-04T23:17:12 I have my Gentoo Linux set to unicode and really wish utf-8 was everywhere. For example my bot outputs utf-8 and playgame.jar reads it as latin-1 and outputs garbage symbols 2010-11-04T23:17:44 *** narnach has quit IRC (Ping timeout: 260 seconds) 2010-11-04T23:17:45 I want to force everyone to use ascii :P 2010-11-04T23:17:47 they have these nice 'airplane' and 'anchor' symbols for fleets and arrows for growth 2010-11-04T23:17:53 haha 2010-11-04T23:18:50 it is better than to write "planet 1: growth 5, ships 100" - and dont tell me i could have written "p1: 100 +5" ^^ 2010-11-04T23:19:41 :P 2010-11-04T23:21:03 So... I'm still trying to speed up my code that simulates every possible move. Somehow this 22*21*20*... stuff doesn't really want to run in under a second for even a small number of turns 2010-11-04T23:22:12 errr... you can't be doing anything with 22! options 2010-11-04T23:22:16 22! = 51090942171709440000 2010-11-04T23:22:18 When I told the code to use no more than 1000ms cpu time it only came up with one totally useless combination of moves until turn 200 2010-11-04T23:22:30 *turn 20, sorry 2010-11-04T23:22:53 *** narnach has joined #aichallenge 2010-11-04T23:23:02 sigh: ...that... is a fairly large number... which i have... taken into acount of course 2010-11-04T23:23:48 it on the same order of magnitude as 2**64 2010-11-04T23:24:08 damn, I need a new integer type in c++ 2010-11-04T23:25:07 *** tapwater has joined #aichallenge 2010-11-04T23:26:14 seriously, i'm now doing a breadth-first-search and only put results on the top of the queue that give me more growth or ships for a given round. i'm currently running tests on the 100 maps to see if the results are 99% identical between a timelimit of 1s and 3min. 2010-11-04T23:26:44 sigh: what is 2**64 never seen that 2010-11-04T23:26:55 2^64 2010-11-04T23:27:07 demerzel_: yeah, it's common in programming languages where ^ is binary xor 2010-11-04T23:27:08 oh weird never seen that notation befoer lol 2010-11-04T23:27:28 *bitwise xor 2010-11-04T23:27:48 yea i'm gonna play with bit shifting if i ever get a stable enough state to test moves on lol 2010-11-04T23:28:09 *** Arthur has joined #aichallenge 2010-11-04T23:28:11 lot of these old school ai progs in c used bitmasks etc tricks to sepeed em up 2010-11-04T23:28:40 demerzel_: I seriously would not worry about that until you have your algorithm down completely 2010-11-04T23:28:43 don't you represent planet ownership in a 32-bit int bitmask? 2010-11-04T23:29:20 *** Arthur has quit IRC (Client Quit) 2010-11-04T23:29:26 i don't know how to do bitmasks in full c compliance tbh 2010-11-04T23:29:36 *** Aaaarg has joined #aichallenge 2010-11-04T23:29:58 Frontier: you do that? 2010-11-04T23:30:21 you're telling me you're trying to do any kind of anything 5.109b times per second? 2010-11-04T23:30:25 sigh: What am I supposed to do? With 22! different combinations my 2GB were full in an instant 2010-11-04T23:30:26 i don't think that will happen 2010-11-04T23:30:45 Frontier: ok fine, you have a memory bottleneck :) 2010-11-04T23:31:29 but to be honest, the with 22! that's an understatement 2010-11-04T23:31:55 demerzel_: 5.109b? 2010-11-04T23:32:08 lol, I hope if I add more and more heuristics to my supposedly "exhaustive" search, it doesn't become all heuristical in the end. 1 second is a short time frame after all. 2010-11-04T23:32:11 err trillion even 2010-11-04T23:32:17 isn't that 21! 2010-11-04T23:32:25 why did this window stop autoscrolling >:( 2010-11-04T23:32:30 22! = 51090942 trillion 2010-11-04T23:33:18 lesson for today: factorial grows fast! 2010-11-04T23:33:38 haha 2010-11-04T23:33:58 do you actually gain anything by bitmasking in c 2010-11-04T23:34:38 i haven't ever bothered with it becaues my day job is C# and i've never had a calc crunch like that before 2010-11-04T23:34:40 you can reduce the amount with an algorithm i invented yesterday: you pair your moves and use the best combination in the set. If you look at planet 3,4,5 you get 6 possible orders. Then you calculate which one is best and only keep that when you add your fourth planet. 2010-11-04T23:34:51 in c that is since C# probably would make a bitmask slower lol 2010-11-04T23:34:55 It will break down the complexity to 2^n 2010-11-04T23:35:16 exponential grows fast too, but not as bad a factorial :) 2010-11-04T23:35:20 so what do you reduce 21! to 2010-11-04T23:35:28 4.000.000 2010-11-04T23:35:32 oh wow that's nice 2010-11-04T23:35:57 isn't that just a/b 2010-11-04T23:36:00 hey guys, i have an algorithms question that i can't figure out 2010-11-04T23:36:02 i have a 2d array of heights 2010-11-04T23:36:03 neighboring heights must respect the following heightA/8=heightB/8 2010-11-04T23:36:05 exception happens only when abs(heightA/8-heightB/8)=1 and either of them %8=0 2010-11-04T23:36:06 i need to change that array somehow, so it respects the rules mentioned 2010-11-04T23:36:38 heightA/8=heightB/8 => heightA=heightB ? 2010-11-04T23:37:31 there is only one problem: if you cannot find a 'best' solution between (1->2->3, 2->3->1 and 1->3->2) you would have to keep them all and gradually get back to 22!. For example one solution gives you 200 ships and 20 growth while another gives you 100 ships and 15 growth. 2010-11-04T23:37:32 sigh: integer division 2010-11-04T23:37:39 ahhh 2010-11-04T23:37:57 *err 100 ships and 25 growth 2010-11-04T23:38:28 sounds like you're way ahead of me on the brute force path lol 2010-11-04T23:38:50 *** McLeopold has joined #aichallenge 2010-11-04T23:39:27 holy shit c bitmask syntax is insane 2010-11-04T23:39:33 amstan: but then all heights have to have the same value for height/8 ? 2010-11-04T23:39:44 sigh: no they don't 2010-11-04T23:39:58 one can be 0 the next one 8, the next one 16 2010-11-04T23:40:22 amstan: ok so it is: heightA%8=heightB%8? 2010-11-04T23:40:23 oh, I see... that's what you meant by exception 2010-11-04T23:40:29 Frontier: no 2010-11-04T23:41:16 here's a valid one: 1,2,3,4,8,9,10,16,17,18 2010-11-04T23:41:30 this is invalid: 1,2,3,9,10 2010-11-04T23:41:49 since 3/8=0, 9/8=1 and neither of them are 8 2010-11-04T23:42:20 now.. the check is easy to implement, using my rules 2010-11-04T23:42:32 but i need to modify an array that i have to respect those rules 2010-11-04T23:42:44 amstan: it really depends on what you want the result to be... I'd start from one square and propogate out 2010-11-04T23:42:45 i can modify it as long as it kinda looks the same 2010-11-04T23:43:20 amstan: but if you do that, the result might be completely different 2010-11-04T23:43:47 oh good i don't need to learn c bitmasking: First rule of bit masks: Don't. Don't unless you have a really bad need to save a tiny amount of memory, and slowing down your code a lot isn't a problem 2010-11-04T23:44:15 I'd find the 8's first. Those are your magical "bridge" numbers. 2010-11-04T23:44:29 demerzel_: depends how you are using them 2010-11-04T23:44:52 says they are inherently nonportable and many compilers ignore them 2010-11-04T23:44:56 *** davidjliu has joined #aichallenge 2010-11-04T23:44:59 McLeopold: but you can change squares so that they are 8s 2010-11-04T23:45:00 i wouldn't chance it 2010-11-04T23:45:19 & in a bit mask is faster than comparing each flag individually in a word for instance 2010-11-04T23:45:27 demerzel_: field[x] vs. mask & (1 << x), it shouldn't be *that* slow 2010-11-04T23:45:45 sigh: what squares? I think I missed something? 2010-11-04T23:45:53 Frontier: and you can compare more than one flag at a time with & 2010-11-04T23:46:16 it really depends on how you are using them demerzel_ 2010-11-04T23:46:26 Frontier: i know but why not just increment pointers - only diff there is the minute memory difference of 23*3 bits 2010-11-04T23:46:40 McLeopold: ah, you arrived in the middle of the converation... amstan had a problem regarding a heighmap 2010-11-04T23:46:56 let me find screenshots.. 2010-11-04T23:47:02 i have no idea what a heightmap is 2010-11-04T23:47:10 demerzel_: 2d array of heights :P 2010-11-04T23:47:20 well yea 2010-11-04T23:47:22 for what? 3d graphics? 2010-11-04T23:47:25 demerzel_: think of a generated landscape 2010-11-04T23:47:32 amstan: is generating random terrain, i think 2010-11-04T23:47:43 i am 2010-11-04T23:47:45 i toyed with 3d java for a while before shooting myself for considering it 2010-11-04T23:48:03 well.. i'm trying to import arbitrary heightmaps in cube2 2010-11-04T23:48:15 amstan: so what kind of transformations are allowed? 2010-11-04T23:48:28 McLeopold: mostly anything, as long as the numbers are close 2010-11-04T23:49:07 oh that engine that runs my sauerbraten 2010-11-04T23:49:09 so for your invalid sequence: 1,2,3,9,10: I can change it to 1,2,3,8,10? 2010-11-04T23:49:20 McLeopold: yes 2010-11-04T23:49:24 amstan: what goes wrong if you fix a single square and propagate the constraints in a breath-first manner? 2010-11-04T23:49:32 in the 2d case 2010-11-04T23:49:43 *breadth-first 2010-11-04T23:49:43 do diagonals count? 2010-11-04T23:49:49 yes 2010-11-04T23:50:36 so you're trying to find a good way to make a boxy looking terrain? 2010-11-04T23:50:46 here's a screenshot showing this: http://pastebin.com/VWGqTgH3 2010-11-04T23:51:06 each of those cubes can be divided in 8 2010-11-04T23:51:20 the heights of each of the corners cannot pass through the cube boundaries 2010-11-04T23:51:39 should I see a picture? 2010-11-04T23:51:47 wait.. 2010-11-04T23:51:48 amstan: that's pastebin 2010-11-04T23:51:49 darn 2010-11-04T23:51:51 http://csclub.uwaterloo.ca/~amstan/upload/screenshot_130824.png 2010-11-04T23:51:53 yeah 2010-11-04T23:51:54 sorry 2010-11-04T23:52:01 will it look like minecraft? 2010-11-04T23:52:09 Frontier: ideally it shouldn't 2010-11-04T23:52:28 haha minecraft 2010-11-04T23:52:35 i still have a server running on a vps 2010-11-04T23:52:42 been waiting for like 2 months for an update 2010-11-04T23:53:54 gnite good luck with your raster 2010-11-04T23:54:01 *** demerzel_ has quit IRC (Quit: Page closed) 2010-11-04T23:56:01 amstan: the more numbers divisible by 8 that exist, the easier it will be? 2010-11-04T23:56:29 i guess so, but then it'll look blocky 2010-11-04T23:56:39 with 45deg angles everywhere 2010-11-04T23:56:53 so, for a start, round everything down to an number divisible by 8? 2010-11-04T23:56:56 ah, interpolated minecraft :P 2010-11-04T23:57:02 we can improve upon that start 2010-11-04T23:58:17 McLeopold: ok.., then what? 2010-11-04T23:58:45 in the picture, when you have some of those stair steps, you want to bump up the lower part so it is smooth? 2010-11-04T23:59:17 McLeopold: pretty much 2010-11-04T23:59:46 McLeopold: that error there looks like a 6,9