2010-09-15T00:00:22 rogue780|server, https://gist.github.com/9cde482d838dab297d0f 2010-09-15T00:01:32 *** fisxoj has joined #aichallenge 2010-09-15T00:01:47 Queue29, thanks 2010-09-15T00:02:01 I still use Gentoo 2010-09-15T00:02:29 is there some gentoo package I need to compile the server maybe? 2010-09-15T00:02:40 dwchandler: the server looks good... thanks for setting it up. 2010-09-15T00:02:44 *** Any_Key has joined #aichallenge 2010-09-15T00:03:19 how i see all gentoo users: http://funroll-loops.info/ 2010-09-15T00:03:35 rogue780: phxbsd.com/tcp.c 2010-09-15T00:03:54 Hey, my server is purple with decals. How'd you know. 2010-09-15T00:03:55 dmj111: np :) 2010-09-15T00:04:10 I found something about testing lisp bots on another server on the forum, any info on how to do that? 2010-09-15T00:04:53 so...I had to replace arc4random() with rand() 2010-09-15T00:04:57 I hope that didn't change much 2010-09-15T00:05:19 I have to have a smoke and a stiff drink before getting into httpd config 2010-09-15T00:05:21 yeah, arc4random is troubling me too 2010-09-15T00:05:29 fisxoj, until earlier today, there was a nice alternative server that we were playing on, where the bot would run on your own pc, but it's down 2010-09-15T00:05:35 McLeopold, replace it with rand() and it compiles fine 2010-09-15T00:05:50 Queue29, aw, now we have to wait for lisp support in the main server, then? 2010-09-15T00:05:58 fisxoj, read through this: http://ai-contest.com/forum/viewtopic.php?f=18&t=424 2010-09-15T00:06:08 hey man, i'm waiting for support for D 2010-09-15T00:06:13 arc4random is available on most Linux, but you may need other/different includes, iirc 2010-09-15T00:06:39 okay rogue, now compile the htdocs and see what you get 2010-09-15T00:07:45 *** rwa has quit IRC (Quit: rwa) 2010-09-15T00:08:18 if someone else gets tcp server up that's cool, but I'm also fine with leaving it up on my server. It's decently fast with an ok connection 2010-09-15T00:08:31 whatever works out 2010-09-15T00:08:56 I'd rather only have one or two, so we don't split between a bunch of servers 2010-09-15T00:09:09 I've enjoyed playing with you all so far, and I want it to continue 2010-09-15T00:09:09 *** rogue780|server| has joined #aichallenge 2010-09-15T00:09:14 McLeopold, in a second. I'm routing my network right now 2010-09-15T00:09:23 dwchandler, the elo rankings are fucked on your tcp server 2010-09-15T00:09:47 rogue780, do you know how to start it? I'm assuming there are command line options 2010-09-15T00:09:59 Queue29: hrm... 2010-09-15T00:10:02 In benze server not coming back up? 2010-09-15T00:10:14 *** JCS^ has quit IRC (Quit: \x00) 2010-09-15T00:10:19 daniel will kick it when he wakes up, I iimagine 2010-09-15T00:10:44 *** rogue780|server has quit IRC (Ping timeout: 245 seconds) 2010-09-15T00:10:48 that would be best, all around 2010-09-15T00:10:55 *** rogue780|server| is now known as rogue780|server 2010-09-15T00:11:08 @later tell danielvf Really apprciate the TCP server. Thanks a bunch. I think it is down right now, though; just FYI. 2010-09-15T00:11:08 bss03: The operation succeeded. 2010-09-15T00:11:45 bss03: is danielvf also dhartmei? 2010-09-15T00:12:03 dhartmei is the one with the server 2010-09-15T00:12:44 *** rwa has joined #aichallenge 2010-09-15T00:13:12 can someone try the tcp server 68.34.234.213 port 9999 ? 2010-09-15T00:13:45 *** Obeleh has joined #aichallenge 2010-09-15T00:14:02 rogue780|server, i could, yes 2010-09-15T00:14:09 *** AndroUser has joined #aichallenge 2010-09-15T00:14:42 . 2010-09-15T00:14:46 . 2010-09-15T00:14:51 my tcpserver is up at 68.34.234.213 port 9999 2010-09-15T00:15:26 can ya'll connect to it? 2010-09-15T00:15:56 ya 2010-09-15T00:15:58 rogue780, yep, i'm in, playing ErrBot_0.0.5 2010-09-15T00:16:08 Queue29, then you better PREPARE TO WIN 2010-09-15T00:16:21 VICTORY IS MINE 2010-09-15T00:16:47 rogue780: can you get the htdocs compiled? 2010-09-15T00:16:54 no, I'm missing atol() 2010-09-15T00:16:56 *** ermau has quit IRC (Quit: ermau) 2010-09-15T00:17:04 looks like I logged in ok. 2010-09-15T00:17:24 yeah! that's what I got. I added stdlib.h and got past that. 2010-09-15T00:17:36 McLeopold, you're looking for #gentoo 2010-09-15T00:17:36 take that ErrBot! 2010-09-15T00:17:41 why mor tcp servers? 2010-09-15T00:17:48 rogue780: I'm not connecting to yours :( 2010-09-15T00:17:58 antimatroid, benzedrine is down 2010-09-15T00:18:28 *** Obeleh has quit IRC (Ping timeout: 265 seconds) 2010-09-15T00:18:29 dwchandler, weird. I've got people playing 2010-09-15T00:18:59 i'm connected to both.. AT THE SAME TIME :O 2010-09-15T00:20:30 *** braddunbar has joined #aichallenge 2010-09-15T00:21:34 I keep getting this: 2010-09-15T00:21:36 oh yeah, sitting in 5th on official leaderboard 2010-09-15T00:21:36 '010.09.15 00:15:20 handle_line: player ErrBot_0.0.5: unexpected line ' 2010-09-15T00:21:36 '010.09.15 00:15:20 handle_line: player ErrBot_0.0.5: unexpected line 'C:\Users\Shawn\Desktop\JAVA_S~1>java ErrBot 2010-09-15T00:21:41 not that that counts for much yet 2010-09-15T00:21:50 *** DirtyKeyboard has quit IRC (Quit: DirtyKeyboard) 2010-09-15T00:21:58 did you open up the http side for rankings? 2010-09-15T00:22:15 Cyndre, what do you mean? 2010-09-15T00:22:27 trying to see the player rankings 2010-09-15T00:22:39 Cyndre, I'm still trying to get that part running 2010-09-15T00:22:55 k, was wondering 2010-09-15T00:24:00 alrighty...stupid bsd programmers 2010-09-15T00:24:07 and by stupid I mean makes life hard for me 2010-09-15T00:24:33 rogue780|server: now you know how it feels to try compiling random linux-only stuff on bsd ;-) 2010-09-15T00:24:52 hey, I code to the ANSI standard. I don't use strlcpy() 2010-09-15T00:24:53 @rankings 2010-09-15T00:24:54 deepblue: Top 10 players: Tracer(2985), deepblue(2938), _Shurf(2917), pdbpdb80(2890), youwillfail(2886), Loks1982(2874), damien(2872), Macuyiko(2791), Jerome(2786), a1k0n(2769) 2010-09-15T00:25:11 except dhartmei writes pretty clean stuff, so it's just a tiny bit to port it 2010-09-15T00:25:26 *** Any_Key has quit IRC (Ping timeout: 264 seconds) 2010-09-15T00:26:06 so what is strlcpy's equivilant? 2010-09-15T00:26:50 #include 2010-09-15T00:26:50 #include 2010-09-15T00:26:50 size_t strlcpy(char *, const char *, size_t); 2010-09-15T00:26:50 * Copy src to string dst of size siz. At most siz-1 characters 2010-09-15T00:26:52 * will be copied. Always NUL terminates (unless siz == 0). 2010-09-15T00:26:54 * Returns strlen(src); if retval >= siz, truncation occurred. 2010-09-15T00:26:56 */ 2010-09-15T00:27:00 size_t 2010-09-15T00:27:02 strlcpy(char *dst, const char *src, size_t siz) 2010-09-15T00:27:04 { 2010-09-15T00:27:06 char *d = dst; 2010-09-15T00:27:08 const char *s = src; 2010-09-15T00:27:10 size_t n = siz; 2010-09-15T00:27:12 /* Copy as many bytes as will fit */ 2010-09-15T00:27:14 if (n != 0 && --n != 0) { 2010-09-15T00:27:16 do { 2010-09-15T00:27:18 if ((*d++ = *s++) == 0) 2010-09-15T00:27:20 break; 2010-09-15T00:27:21 are you kidding me 2010-09-15T00:27:22 } while (--n != 0); 2010-09-15T00:27:24 } 2010-09-15T00:27:26 /* Not enough room in dst, add NUL and traverse rest of src */ 2010-09-15T00:27:30 you couldn't have put that on codepad or something? 2010-09-15T00:27:30 if (n == 0) { 2010-09-15T00:27:32 if (siz != 0) 2010-09-15T00:27:34 *d = '\0'; /* NUL-terminate dst */ 2010-09-15T00:27:36 while (*s++) 2010-09-15T00:27:38 ; 2010-09-15T00:27:40 } 2010-09-15T00:27:42 return(s - src - 1); /* count does not include NUL */ 2010-09-15T00:27:44 } 2010-09-15T00:27:45 * bss03 kicks Queue29. 2010-09-15T00:27:47 is benzedrine hanging ? 2010-09-15T00:27:48 he asked what the equivalent was :( 2010-09-15T00:27:54 bartwe: looks like it 2010-09-15T00:27:56 McLeopold, strcpy 2010-09-15T00:27:59 agh, pastebin.org please 2010-09-15T00:28:01 from here, anyway http://ubuntuforums.org/showthread.php?t=128567 2010-09-15T00:28:36 *** deepblue has quit IRC (Quit: Page closed) 2010-09-15T00:28:37 pastebin.ca ftw 2010-09-15T00:29:06 *** AndroUser has quit IRC (Ping timeout: 265 seconds) 2010-09-15T00:29:22 so...am I supposed to run the webserver portion with an apache backend? 2010-09-15T00:30:25 *** fisxoj has quit IRC (Quit: Ex-Chat) 2010-09-15T00:32:28 ok. i'm guessing apache2 with mod_cgi 2010-09-15T00:33:45 looks like each directory gets mapped to each cgi program 2010-09-15T00:33:54 hmm. I don't know how "ratings.txt" is produced 2010-09-15T00:33:55 *** dhartmei has quit IRC (Quit: Leaving) 2010-09-15T00:34:27 I don't have a ratings.txt, and grep doesn't show me anything creating one 2010-09-15T00:35:37 touch ratings.txt 2010-09-15T00:35:38 dwchandler, from your tcpserver directory 2010-09-15T00:35:41 touch ratings.txt 2010-09-15T00:35:54 pretend you're pedobear and it's justin bieber 2010-09-15T00:36:24 *** Obeleh has joined #aichallenge 2010-09-15T00:36:26 ok...that may have gone too far 2010-09-15T00:39:10 rogue780|server: I did that a bit ago, but I don't see any code that *writes* to results.txt, and it's still 0 size 2010-09-15T00:39:22 dwchandler, same here 2010-09-15T00:39:51 there must be something that reads results.pgn and writes results.txt, but bugger if I know 2010-09-15T00:40:03 ok...anyone know how to enable an apache mod on ubuntu? i tried sudo ln -s ./cgi.load ../mods-enabled/cgi.load, but cgi.load shows up in a funny red color now that makes me think it didn't work 2010-09-15T00:40:22 *** Naith has joined #aichallenge 2010-09-15T00:40:49 rogue780|server: getratings will die without ratings.txt, fyi :/ 2010-09-15T00:41:34 *** Janzert1 has joined #aichallenge 2010-09-15T00:41:37 rogue780|server, red just means the file type is symlink 2010-09-15T00:41:39 *** Janzert has quit IRC (Read error: Connection reset by peer) 2010-09-15T00:42:09 ... waiting apache2: Syntax error on line 204 of /etc/apache2/apache2.conf: Could not open configuration file /etc/apache2/mods-enabled/cgi.load: Too many levels of symbolic links 2010-09-15T00:42:21 *** Janzert1 is now known as Janzert 2010-09-15T00:43:56 i can't find results.pgn 2010-09-15T00:44:39 server writes it in current directory 2010-09-15T00:44:59 not over here. it doesn't exist 2010-09-15T00:45:39 odd. writes it out here ok 2010-09-15T00:46:07 hey, benzedrine.cx is up 2010-09-15T00:46:14 yay! 2010-09-15T00:47:39 *** mark_jn has joined #aichallenge 2010-09-15T00:47:54 *** mark_jn has left #aichallenge 2010-09-15T00:48:34 hey, anyone know how to get apache to run cgi files instead of offering to download them? 2010-09-15T00:48:43 anyone know who bartwe is? 2010-09-15T00:49:11 benzedine is up again 2010-09-15T00:49:14 His real name is Leslie Simmons, his friends call him "Lez" 2010-09-15T00:49:32 durrrr, sorry about that echo 2010-09-15T00:50:10 who runs benzedine? 2010-09-15T00:50:15 now that benzedrine is back, I got new server.tar.gz and it has a script to make ratings.txt! but now I don't need it :) 2010-09-15T00:50:15 rogue780|server: you have to enable execCGI for the directory they're in... I think 2010-09-15T00:50:25 rogue780|server: dhartmei 2010-09-15T00:50:27 Options +ExecCGI 2010-09-15T00:50:59 ^what they said 2010-09-15T00:52:12 McLeopold, other_donkey, here's my 000-default file...http://pastebin.ca/1941105 2010-09-15T00:52:24 I can get a directory listing of /home/shawn/tcpserver/htdocs 2010-09-15T00:53:01 but when I try and do http://localhost/getplayer?rogue780 it just tries and serve me up getplayer as a download 2010-09-15T00:53:06 *** HakanD has quit IRC (Ping timeout: 252 seconds) 2010-09-15T00:53:14 I have an AddHandler cgi-script .py in my server 2010-09-15T00:53:34 But I'd have to lookup without file extensions 2010-09-15T00:55:24 rogue780|server: is the script chmod'd executable? 2010-09-15T00:55:44 yes 2010-09-15T00:57:14 if you got getgame.c to compile, could you pastebin it? 2010-09-15T00:57:31 *** mark_jn has joined #aichallenge 2010-09-15T00:58:08 Maybe you need a ScriptAlias 2010-09-15T00:58:32 McLeopold, http://pastebin.ca/1941111 2010-09-15T00:58:57 gah, rate.sh uses readpgn, which is missing :( 2010-09-15T00:59:24 McLeopold, this is strlcpy.h http://pastebin.ca/1941112 2010-09-15T00:59:33 dwchandler, where did you find rate.sh? 2010-09-15T01:00:16 I'm just about to make my own server and viewer setup 2010-09-15T01:00:20 new server.tar.gz from benzedrine 2010-09-15T01:00:23 this is getting ridiculous 2010-09-15T01:01:18 ah, nothing like getting alpha releases to run 2010-09-15T01:02:09 well, I just spent the last few hours getting this server half-way setup instead of working on my bot :\ 2010-09-15T01:02:53 canvas.c needed it too 2010-09-15T01:03:12 McLeopold, all I did was include my strlcpy.h and canvas.c compiled 2010-09-15T01:03:22 yes, it already worked 2010-09-15T01:03:36 and by "my" I mean the one that Queue29 linked to from ubuntu forums 2010-09-15T01:04:04 whoa, is apache.org down or my connection just stupid? 2010-09-15T01:04:29 I was on it about 5 minutes ago 2010-09-15T01:04:48 all signs point to stupid connection 2010-09-15T01:05:05 *** hornairs has quit IRC (Quit: hornairs) 2010-09-15T01:05:55 *** hornairs has joined #aichallenge 2010-09-15T01:05:55 rogue780|server: did you add a ScriptAlias for the cgi directory? 2010-09-15T01:06:14 other_donkey, not sure what that is 2010-09-15T01:06:24 http://httpd.apache.org/docs/current/mod/mod_alias.html 2010-09-15T01:06:32 towards the bottom 2010-09-15T01:06:41 hey, whoever is on my server, I'm going to be bringing it down in a once your game is over and try the new code 2010-09-15T01:07:03 *** Naith has quit IRC (Quit: Naith) 2010-09-15T01:07:48 What is the situation with tcp bot servers? 2010-09-15T01:07:57 Do I need to modify my client? 2010-09-15T01:08:09 greghaynes: could you add an address field? 2010-09-15T01:08:37 I've been editing the source to change servers 2010-09-15T01:08:37 Sure 2010-09-15T01:08:41 heh 2010-09-15T01:08:51 rogue780: Which address was you? 2010-09-15T01:08:54 *** hellman has quit IRC (Read error: Connection reset by peer) 2010-09-15T01:09:34 rogue780: Or more specifically, what address are the most people connected to? (So I can connect there.) 2010-09-15T01:09:36 *** mark_jn has quit IRC (Quit: leaving) 2010-09-15T01:09:38 rogue780|server: if you get the later server.tar.gz, be aware to touch users.txt or nobody can play 2010-09-15T01:10:36 looks like readpgn is part of bayeselo stuff, which I haven't downloaded yet 2010-09-15T01:12:18 other_donkey: Ok, try now 2010-09-15T01:12:48 dwchandler.networkninjas.net and rogue780.networkninjas.net should work for those user's respective servers 2010-09-15T01:13:24 *** Janzert1 has joined #aichallenge 2010-09-15T01:13:25 Yep, works 2010-09-15T01:13:26 *** Janzert has quit IRC (Ping timeout: 264 seconds) 2010-09-15T01:18:27 Has the C++ package been updated recently? I haven't done anything for a while :) 2010-09-15T01:20:48 well, I'm done for the night. feel free to play on my server, but there won't be any stats until I get things figured out. tomorrow I'm going to focus on an updated bot. g'night 2010-09-15T01:21:40 *** Any_Key has joined #aichallenge 2010-09-15T01:22:23 greghaynes: awesome, works perfectly 2010-09-15T01:22:44 rogue780|server: thanks for the server 2010-09-15T01:23:18 *** Yoshi-TS4 has joined #aichallenge 2010-09-15T01:23:50 dwchandler: is your server up? 2010-09-15T01:25:29 should be 2010-09-15T01:25:35 1 connection 2010-09-15T01:26:04 I just connected and playing 2010-09-15T01:29:46 *** Yoshi-TS4 has quit IRC (Ping timeout: 240 seconds) 2010-09-15T01:29:54 *** braddunbar has quit IRC (Ping timeout: 240 seconds) 2010-09-15T01:30:34 dwchandler: nm, my dns is suck, dwchandler.networkninjas.net hasn't propagated to my local dns server yet 2010-09-15T01:31:16 dwchandler: also, your bot just totally pwnd mine 2010-09-15T01:32:40 heh 2010-09-15T01:33:13 at least I can beat error_bot 2010-09-15T01:33:22 140.99.51.114 == stilyagin.com 2010-09-15T01:33:34 *** McLeopold has quit IRC (Quit: Leaving) 2010-09-15T01:35:40 *** caytchen has joined #aichallenge 2010-09-15T01:36:19 cool, that saves me from adding strange dns entries to my zone 2010-09-15T01:38:06 :) 2010-09-15T01:38:08 *** braddunbar has joined #aichallenge 2010-09-15T01:38:59 dwchandler: are you Queue29_DBot? 2010-09-15T01:39:10 I can't get to the bayeselo site, but google pointed me to a third-party directory on ... wait for it ... ai-contest.com :) 2010-09-15T01:39:37 * dwchandler can has ratings 2010-09-15T01:39:48 mm 2010-09-15T01:39:58 other_donkey: no, I'm dwc-### 2010-09-15T01:49:05 bartwe6 i can't beat but 7 i can on the off chance aha 2010-09-15T01:50:26 i really wish they'd just set a cap on fleet numbers already if they're going to do that 2010-09-15T01:50:30 makes the problem way harder 2010-09-15T01:50:56 at least for what i'm intending to do 2010-09-15T01:52:56 other_donkey, I believe Queue29_DBot belongs to Queue29 2010-09-15T01:52:56 *** syntaxglitch has quit IRC (Ping timeout: 272 seconds) 2010-09-15T01:53:19 hi 2010-09-15T01:53:26 yeah, just realized that 2010-09-15T01:53:33 other_donkey, as far as err_bot goes, the one running is seriously flawed. yesterday's was much better and tomorrow's will be better again I hope 2010-09-15T01:53:37 anti: 7 == 6 unless i introduced a bug 2010-09-15T01:53:47 anyawy, off to bed for real 2010-09-15T01:53:57 7 was a rewrite to be allocation free so that the main servers would accept it 2010-09-15T01:55:42 anyone using a* or redblack so far ? 2010-09-15T01:56:01 *** jaspervdj has joined #aichallenge 2010-09-15T01:56:38 bss03: I'm pulling your request, I'll then ping the organizers 2010-09-15T01:57:31 i'm considering a*, but need to reduce the search space heaps more first 2010-09-15T01:59:02 What do you guys need path algorithms for? 2010-09-15T01:59:23 a* is a way to search a statespace 2010-09-15T01:59:35 g'nite 2010-09-15T01:59:47 *** dhaval has joined #aichallenge 2010-09-15T01:59:50 creat e a heuristic of closeness to yout 'goal' and off you go 2010-09-15T01:59:56 a* search as in the algorithm for searching trees yeah? 2010-09-15T02:00:11 Yeah, any connected graph. 2010-09-15T02:01:02 nodes in the tree are possilbe game states 2010-09-15T02:01:13 edges are possible fleet movements 2010-09-15T02:01:41 now you need a way to generate a small enough number of edges 2010-09-15T02:01:46 thats a fat graph 2010-09-15T02:01:49 and a scoring function for a state 2010-09-15T02:02:07 then use a* to dynamicly deepen the graph 2010-09-15T02:02:09 first remove any dominated strategies, then do waht you can do pick your favourite "pseudo" optimal ones 2010-09-15T02:02:23 bartwe: your newest bot owns very hard 2010-09-15T02:02:32 Baphomet: which one ? 2010-09-15T02:02:38 7.6.3 2010-09-15T02:02:49 hasn;t beaten .6 consistently yet 2010-09-15T02:03:28 antimatroid: im playing you right now, your going down buddy!! 2010-09-15T02:03:34 or not. 2010-09-15T02:03:37 Baphomet: you are actually doing quite well, considering how you thought you would the other day 2010-09-15T02:03:38 :D 2010-09-15T02:03:47 trying to add resilience to some tactics that do me in, which makes it weaker against the older, more focused version 2010-09-15T02:03:50 thanks 2010-09-15T02:04:23 i do like your support of the front line 2010-09-15T02:04:25 did benzedrine loose its reocrd db today? 2010-09-15T02:04:28 i haven't even touched that yet 2010-09-15T02:04:40 yeah my elo reset 2010-09-15T02:04:47 then i got pounded by bartwe 2010-09-15T02:05:09 the tcp server apparently lets you issue invalid orders? 2010-09-15T02:05:14 benzedrine was stuck this morning 2010-09-15T02:05:18 You issued an invalid order: src 20, dst 21, ships 3109 2010-09-15T02:05:26 You WIN against dhartmei 2010-09-15T02:05:31 har 2010-09-15T02:05:43 now that is a way to win 2010-09-15T02:05:51 hmmm, i've just set my program up to never ever make an illegal move 2010-09-15T02:06:03 and i'm being maticulous about a timer after what happened to me last year 2010-09-15T02:06:10 well, earlier this yera* 2010-09-15T02:06:25 my bot makes tones of illegal movies 2010-09-15T02:06:29 moves* 2010-09-15T02:06:32 *** caytchen has quit IRC (Ping timeout: 276 seconds) 2010-09-15T02:06:44 and i can tell when im loosing by how many illegals 2010-09-15T02:06:58 ha, fair enough, you will lose those games on the main server 2010-09-15T02:07:27 it doesnt actually, only illegal ones it does is send 0 ships 2010-09-15T02:07:37 *** kurige has joined #aichallenge 2010-09-15T02:08:35 oh okay, i think there were still intending to make that an illegal move, may not have happened 2010-09-15T02:08:50 I haven't been able to submit my new bot for two days. No compilation error or email. Is this a known problem, or am I just "special"? 2010-09-15T02:09:04 kurige: it should be working again now 2010-09-15T02:09:16 i got a bot up last night (4pm here) and it's had some games 2010-09-15T02:09:26 (i'm "youwillfail" on the main server this time around) 2010-09-15T02:10:11 i swear my bot is worse now :( 2010-09-15T02:10:14 hmm... not for me. I get the "Success!" page, and then nothing happens when I check back at 10 min intervals. 2010-09-15T02:10:48 Nobody else is having problems with this? 2010-09-15T02:11:14 Baphomet: i've made many a changes that just make it flop 2010-09-15T02:11:31 kurige: it usualyl takes a while to get my first game 2010-09-15T02:11:42 just upload it and leave it as is for a few hours at least 2010-09-15T02:11:53 antimatroid: but does it show up in "Recent Submissions" *before* you get a game? 2010-09-15T02:12:04 hmmm, not sure 2010-09-15T02:12:25 danielvf: you about? 2010-09-15T02:12:29 antimatroid: hmm... well, I'll leave it alone for 24hrs and see what happens. :( 2010-09-15T02:12:51 amstan: The haskell package is more or less done 2010-09-15T02:12:57 *** Palmik has joined #aichallenge 2010-09-15T02:13:01 antimatroid: hehe, that's a pretty pessimistic name 2010-09-15T02:13:02 amstan: What should be done to make it official? 2010-09-15T02:13:12 kurige: have you tried the unofficial tcp server? 2010-09-15T02:13:17 you will get more games and against better bots anyway 2010-09-15T02:13:35 *** vizier has joined #aichallenge 2010-09-15T02:13:54 antimatroid: I saw it, but never got around to trying it. Does the TCP stuff work on cygwin? 2010-09-15T02:14:12 *** preetum has joined #aichallenge 2010-09-15T02:14:15 it should, i couldn't get cygwin to pick up gcc 2010-09-15T02:14:22 so i'm running with some java class file 2010-09-15T02:14:35 i can't believe top bot on main server is java 2010-09-15T02:14:45 lol 2010-09-15T02:14:54 antimatroid: I saw the java one too... I think I'll just go with that. 2010-09-15T02:14:59 yeah, top bot keeps changing, though 2010-09-15T02:15:07 my bot hasn't played any games since it was uploaded 24 hours ago. Is the engine still having problems, or is this normal? 2010-09-15T02:15:13 it was a python one earlier 2010-09-15T02:15:24 *** tapwater has quit IRC (Quit: tapwater) 2010-09-15T02:15:25 i have had games in the last 24 hours, not sure why you guys aren't get any 2010-09-15T02:15:41 preetum: this doesn't help with your problem, but I'm curious: does your submission show up in your "recent submissions" list? 2010-09-15T02:16:08 kurige: did you get any emails about compilation success? 2010-09-15T02:16:18 i got one last night when i submitted 2010-09-15T02:16:21 kurige: yes, it shows up as successfully compiled (http://www.ai-contest.com/profile.php?user_id=6816) 2010-09-15T02:16:29 *** HakanD has joined #aichallenge 2010-09-15T02:16:40 antimatroid: no. in the past I've only gotten emails for compilation errors. I think the emails are just really, really intermittent. 2010-09-15T02:17:22 *** Any_Key has quit IRC (Ping timeout: 265 seconds) 2010-09-15T02:17:25 prettum: my problem is worse, then. :( Mine won't even show up, let alone get any games. 2010-09-15T02:17:28 *** dhartmei has joined #aichallenge 2010-09-15T02:17:41 i saw someone say somewhere that the server was sending emails for successes and not errors, and that they were then going to fix that, but may not have 2010-09-15T02:17:46 *** td123 has joined #aichallenge 2010-09-15T02:18:04 antimatroid: ironically google flagged them as spam. I thought they had fixed that as well. 2010-09-15T02:18:14 * kurige goes to check his spam folder 2010-09-15T02:18:37 i think the google name was more of a, this looks good for google and encourages more people to participate deal than any kind of actual support 2010-09-15T02:20:09 *** Any_Key has joined #aichallenge 2010-09-15T02:20:22 already more than 1000 bots up, i wonder how many non-starter package bots we'll end up with in 2 months time 2010-09-15T02:20:51 *** jared has quit IRC (Remote host closed the connection) 2010-09-15T02:21:56 antimatroid: what's the command-line to initiate a TCP battle? Is everyone on the Benzedrine server? 2010-09-15T02:22:42 java TCP "213.3.30.106" "9999" "antimatroid2.0" ".\bot.exe" "100000" 2010-09-15T02:22:44 i use that 2010-09-15T02:22:51 the last argument is some retarded useless thing it wants 2010-09-15T02:23:11 and the third one is whatever name I want to give my bot? 2010-09-15T02:23:16 yep 2010-09-15T02:23:21 thanks muchos 2010-09-15T02:23:35 and fix up location of your bot file 2010-09-15T02:23:57 hey, i beat bartwe6 2010-09-15T02:24:14 is it ok that i use "1000" not "100000"? 2010-09-15T02:24:18 nice 2010-09-15T02:24:25 yeah 2010-09-15T02:24:31 i just chucked wahtever there 2010-09-15T02:25:24 *** zaph has joined #aichallenge 2010-09-15T02:25:43 Not too familiar with Java. It's throwing a NoClassDefFoundError: TCP 2010-09-15T02:26:09 thought just running "java TCP " would do it when I'm in the same folder as the java file. 2010-09-15T02:26:20 pm me your email address, i'll forward you a .class file to put in the folder 2010-09-15T02:26:39 i had that done for me and it did the trick 2010-09-15T02:26:44 same with me for baphomet 2010-09-15T02:27:05 thanks man 2010-09-15T02:27:54 done 2010-09-15T02:28:09 just sharing the love 2010-09-15T02:28:15 ai comps are so damn fun 2010-09-15T02:28:29 although very detrimental to my school work and sleep 2010-09-15T02:28:52 ho ho ho ho! It worked! 2010-09-15T02:28:59 aaaaaaand... "unknown exception" 2010-09-15T02:29:06 hahaha. well, at least it connected! 2010-09-15T02:29:13 hmmm, i've not had that error yet 2010-09-15T02:29:14 thanks again 2010-09-15T02:30:14 Yeah, it connects, starts a game, then barfs. Me thinks it's probably something wrong on my end. 2010-09-15T02:30:29 *** HakanD has quit IRC (Quit: Page closed) 2010-09-15T02:31:30 *** zaph has quit IRC (Quit: Bye) 2010-09-15T02:32:39 Hrm, the spec says that if you run out of ships you lose, but that's not entirely true. 2010-09-15T02:33:02 *** EvGenius has joined #aichallenge 2010-09-15T02:33:13 If, after combat resolution you control no fleets and one planet with zero population and non-zero growth, you don't lose. 2010-09-15T02:33:37 I had a bot hold me in that state for nearly 100 turns day before yesterday, IIRC. 2010-09-15T02:34:42 So, I think it is if you have no ships AND no growth you lose. 2010-09-15T02:34:48 antimatroid: I was trying to run a bot I'd compiled in cygwin from a DOS prompt. Silly me. Thanks again for all your help. I'm "kurige". 2010-09-15T02:34:57 how much memory are you guys using? 2010-09-15T02:35:04 i'm still only using like 1.12mb 2010-09-15T02:35:08 i might have to start storing more 2010-09-15T02:35:08 Not much. 2010-09-15T02:35:42 i can't find anything that works better and requires more memory 2010-09-15T02:35:43 although i guess this leaves room for a big phat tree if one can work out how to effectively use one 2010-09-15T02:36:36 *** O1athe has joined #aichallenge 2010-09-15T02:36:41 *** Olathe has quit IRC (Read error: Connection reset by peer) 2010-09-15T02:38:49 *** EvGenius has left #aichallenge 2010-09-15T02:40:13 i had to redo but to use less memory 2010-09-15T02:40:16 *** Yoshi-TS4 has joined #aichallenge 2010-09-15T02:41:01 *** shawnshekari has joined #aichallenge 2010-09-15T02:41:13 redo what? 2010-09-15T02:41:13 o/ all 2010-09-15T02:41:14 *** jukkerknott has joined #aichallenge 2010-09-15T02:41:20 i redid most of planetwars 2010-09-15T02:41:35 made my life so much easier once i was done 2010-09-15T02:41:45 antimatroid: a* is only used for pathing, right? It's not some substitude for minimax or alpha-beta, right? 2010-09-15T02:42:05 bss03: i have no idea about formal stuff for cs 2010-09-15T02:42:16 i'm final year undergrad student 2010-09-15T02:42:30 and don't really study cs, more maths and economics 2010-09-15T02:43:01 bss03: and wath is a 'path' in a decision tree ? 2010-09-15T02:43:10 *** otherAntimatroid has joined #aichallenge 2010-09-15T02:43:11 i try to think of something that will work, then work out a way that seems efficient to do it 2010-09-15T02:43:12 I keep having these "aha" moments about how to make my bot so much better, then it turns out to be worse 2010-09-15T02:43:14 *** amstan has quit IRC (Ping timeout: 245 seconds) 2010-09-15T02:43:15 antimatroid: At least, from what I learned as an UG and what I'm reading on WP, it doesn't look like you can use it on game trees exactly. 2010-09-15T02:43:35 i think you can on zero sum 2010-09-15T02:43:37 which this game is 2010-09-15T02:43:49 With pathing, your are looking for edges that sum up to a small value. With game tree searching you are looking for a node with high score. 2010-09-15T02:43:53 *** HakanD has joined #aichallenge 2010-09-15T02:44:10 a* search is just with alpha-beta pruning yeah? 2010-09-15T02:44:31 i'm pretty sure it just has a need little trick for storing alpha and beta values at decisions nodes to test whether it's dominated or not 2010-09-15T02:44:34 but i could be wrong 2010-09-15T02:44:38 No, a* is basically dijstras + optimistic heuristic. 2010-09-15T02:44:46 okay, i have no idea then :P 2010-09-15T02:45:58 montalow ? 2010-09-15T02:46:37 Still, I don't know what optimizations you can do on zero-sum games. Game theory is relatively limited in our UG program. 2010-09-15T02:47:32 *** Appleman1234 has quit IRC (Ping timeout: 255 seconds) 2010-09-15T02:48:01 *** Appleman1234 has joined #aichallenge 2010-09-15T02:48:19 UG? 2010-09-15T02:48:39 i'm working on a game theory library in c++, but i don't think any of those things will be very useful 2010-09-15T02:48:43 *** dufflebunk has joined #aichallenge 2010-09-15T02:48:59 UG = undergraduate. 2010-09-15T02:49:17 oh, yeah it's pretty much nonexistent where I am 2010-09-15T02:49:31 i just picked up myerson and tiroles texts and started reading 2010-09-15T02:49:50 i did have a crappy intro course in the eco department at my uni first year, but they didn't even use calculus ffs 2010-09-15T02:49:55 any git users here? 2010-09-15T02:51:09 rwa: I use git. 2010-09-15T02:51:15 rwa: You gots a problem? 2010-09-15T02:51:43 If i have a working directory with changes in it already, how can I commit those to a branch? Right now I'm on the trunk 2010-09-15T02:52:21 (git branch $new_branch_name) then ... 2010-09-15T02:52:22 I think I'm going to need to do this a lot for ideas that don't work out... :p 2010-09-15T02:52:26 (git commit) 2010-09-15T02:53:22 Creating a new branch is basically just making another link to the current HEAD, so you can do it with a dirty tree, normally. 2010-09-15T02:53:53 Oh, wait. I gues I forgot a step. 2010-09-15T02:54:09 Use (git checkout -b $new_branch) instead of git-branch. 2010-09-15T02:54:29 *** jukkerknott has quit IRC (Ping timeout: 252 seconds) 2010-09-15T02:54:33 yeah, that put my changes in master... no biggie 2010-09-15T02:55:01 Well, now you have a branch that is your old master. It should have the latest commit. 2010-09-15T02:55:17 *** drg has quit IRC (Quit: Page closed) 2010-09-15T02:55:22 You can use (git branch -m $old_name $new_name) to rename stuff, if you want. 2010-09-15T02:55:38 should -> shouldn't. 2010-09-15T02:55:54 Gah. 2010-09-15T02:56:59 You should now have: ... history ... -> old_head (called $new_branch) -> now_head (called master). But, you can use git branch -m to rename stuff if you want. 2010-09-15T02:57:06 the "checkout" syntax freaks me out - I think of that as pulling stuff out of the repo and into my wd. 2010-09-15T02:57:28 I was hesitant to do that with changes in the wd 2010-09-15T02:57:32 That's sort of what it does, but it is kinda smart about it. 2010-09-15T02:57:48 It is possible to use checkout to blow away your changes. 2010-09-15T02:58:19 But, when it is being used to swtich banches it tries to preserve your changes, and fails if it can't. 2010-09-15T02:59:15 (git checkout HEAD -- file_name directory/another_file) will revert your changes on some files, but ... 2010-09-15T02:59:45 (git checkout -b new_branch) will only work if it can preserve your changes. Sometimes git is weird for historical reasons. :P 2010-09-15T03:01:16 *** Zeiris has joined #aichallenge 2010-09-15T03:02:48 *** kurige has quit IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.9/20100824153629]) 2010-09-15T03:07:50 *** aerique has joined #aichallenge 2010-09-15T03:07:53 *** mega1 has joined #aichallenge 2010-09-15T03:08:01 So, did they decide to add turns_remaining to the information we get from the server? 2010-09-15T03:08:32 bartwe: http://www.benzedrine.cx/planetwars/canvas?game_id=1284533567|Baphomet-ai-v5|bartwe.7.7.1 lol 2010-09-15T03:08:40 I think the current proposal was to have something like "go # : turns_remaining 200"... 2010-09-15T03:09:51 Baphomet: That's... interesting. 2010-09-15T03:09:57 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-09-15T03:10:44 Baphomet: That man may be a little broken -- that may actually be optimal play. 2010-09-15T03:11:17 Baphomet: If you don't protect your home planet, you'll lose. It takes too long or too much investment to take another world. 2010-09-15T03:11:44 bss03: you can capture the 20-ship neutral while protecting the home planet 2010-09-15T03:12:03 sigh: I don't think so. 2010-09-15T03:12:07 sigh: That takes 21 guys. 2010-09-15T03:12:27 sigh: Which would but bartwe low enough that a suicide force (all the ships on the planet) would take him. 2010-09-15T03:12:29 you have 25 to play with 2010-09-15T03:12:40 *26 2010-09-15T03:12:45 Looks like only 20 to me. 2010-09-15T03:12:49 *** preetum has quit IRC (Ping timeout: 252 seconds) 2010-09-15T03:12:54 sorry, 25 is right 2010-09-15T03:14:22 if red sent 100 on first move and blue did nothing, then blue would have 25 ships after the attack fleet reached his planet 2010-09-15T03:14:22 *** apophis_ has joined #aichallenge 2010-09-15T03:14:22 since he only needs 0, blue has 25 ships to play with 2010-09-15T03:14:22 sigh: 'k. I'll assume you are right. That's still a pretty slim margin (25 - 21 = 4) to work with. 2010-09-15T03:14:31 :) 2010-09-15T03:14:34 yes, but there is zero danger 2010-09-15T03:15:00 red's only option is to attack the 20-ship neutral or take the other 20-ship neutral 2010-09-15T03:15:21 Sure there is. You send 21 guys to one of the neutral planets, then the next turn the other guy sends 6. 2010-09-15T03:15:39 *** rogue780 has quit IRC (Read error: Connection reset by peer) 2010-09-15T03:15:46 Danger lurks in the heart of all men. :P 2010-09-15T03:15:53 baphomet: neat match. you're defeatable there, but he doesn't take the opening. 2010-09-15T03:15:53  2010-09-15T03:15:53 hmmm... yeah 2010-09-15T03:15:53 Or, I guess 7, rather. 2010-09-15T03:16:02 neutral stealing :) 2010-09-15T03:16:07 i hate that tactic, beats my bots sofar ;p 2010-09-15T03:16:18 ok, now I'm not so sure 2010-09-15T03:16:25 Beats the CRAP out of my bots. and deepblue does it *all the time*.! 2010-09-15T03:16:26 what the optimal strategy is 2010-09-15T03:16:31 * bss03 curses deepblue. 2010-09-15T03:16:34 *** Guest7327 has quit IRC (Ping timeout: 245 seconds) 2010-09-15T03:19:06 *** shawnshekari has quit IRC (Quit: Page closed) 2010-09-15T03:19:39 sigh: who says there is an optimal strategy ? 2010-09-15T03:19:56 what is the best strategy for rock/paper/siccors 2010-09-15T03:20:04 thats the burty of this 2010-09-15T03:20:06 scissors 2010-09-15T03:20:11 buety* 2010-09-15T03:20:13 woops 2010-09-15T03:20:20 beauty 2010-09-15T03:20:24 omfg i can has typos? 2010-09-15T03:20:36 phonetic all the way ;) 2010-09-15T03:20:39 :D 2010-09-15T03:21:23 i'm going to add heaps to my bot over the weekend, hopefully i can consistently own you people then :) 2010-09-15T03:21:34 although i suspect others will be intending the same 2010-09-15T03:21:40 (sure)* 2010-09-15T03:21:54 but i'm +10GMT, so I sort of get a head start 2010-09-15T03:22:02 not on me 2010-09-15T03:22:11 yes, but i started beating you over night :P 2010-09-15T03:22:32 haha 2010-09-15T03:22:40 and got you on the main site earlier (under youwillfail) 2010-09-15T03:22:49 ah, that's you 2010-09-15T03:22:53 :) 2010-09-15T03:23:00 beutifull failure 2010-09-15T03:23:04 my bot on the main site is one from 5 days ago 2010-09-15T03:23:08 http://www.benzedrine.cx/planetwars/canvas?game_id=1284534595|bartwe.7.3|bartwe.7.8 2010-09-15T03:23:17 7.8 was so strong after basicly losing 2010-09-15T03:23:19 yeah, well i had you on tcp last night after you left too :P 2010-09-15T03:23:24 still loses ofcourse 2010-09-15T03:23:26 but doesn't mean anything yet anyway 2010-09-15T03:23:34 :) 2010-09-15T03:23:47 well done (unless it was just my bot timing out) 2010-09-15T03:23:48 i think wtfbbq updated their 2010-09-15T03:23:49 bot 2010-09-15T03:23:57 i was consistently caning that last night in the end 2010-09-15T03:24:08 *** Theobon has quit IRC (Quit: Leaving.) 2010-09-15T03:25:21 it so hard at times to spot the 'losing move' 2010-09-15T03:25:45 what do you think is the point a game is lost ? 2010-09-15T03:26:03 capturing the 44 ship neutral 2010-09-15T03:26:06 *** rogue780 has joined #aichallenge 2010-09-15T03:26:06 i've come back and lost from infront 2010-09-15T03:26:22 *** rogue780 has quit IRC (Read error: Connection reset by peer) 2010-09-15T03:28:43 *** retybok has joined #aichallenge 2010-09-15T03:29:40 retybol: great bot ! 2010-09-15T03:29:53 thanks! 2010-09-15T03:30:08 I'm on the verge of a complete rewrite, I've reached the limits of the current model 2010-09-15T03:30:33 whats the 'current model' ? 2010-09-15T03:30:45 I treat each of my planets independently 2010-09-15T03:31:19 it's great because it's simple, but then I can't make coordinated attacks 2010-09-15T03:31:21 retybok: welcome to the party, i did that last weekend, it was horrible but worth it 2010-09-15T03:31:30 also I need a better risk model 2010-09-15T03:33:08 I'm still amazed to be 25th of the general contest with such a simple bot :) 2010-09-15T03:33:24 that's the beauty of starting early, I guess. Things will get a lot more difficult in a month or so 2010-09-15T03:33:46 Also I guess diversity will fall pretty fast 2010-09-15T03:34:14 i'm really starting to wish there was a 1 fleet/move limit 2010-09-15T03:34:29 drewrobb: why??? 2010-09-15T03:34:31 *** criis has joined #aichallenge 2010-09-15T03:34:42 McLeopold.5 looks fearsome 2010-09-15T03:34:49 *** iFire has joined #aichallenge 2010-09-15T03:34:56 too much complexity 2010-09-15T03:35:09 drewrobb: that's the beauty of it :-D 2010-09-15T03:35:13 *** Palmik_ has joined #aichallenge 2010-09-15T03:35:21 rwa: that's because deepblue isn't playing :-D 2010-09-15T03:35:44 *** sigh has quit IRC (Remote host closed the connection) 2010-09-15T03:35:49 nah, beauty is a balance between complexity and tractability 2010-09-15T03:36:18 beauty is in the eye of the beholder 2010-09-15T03:36:29 ...consequently, my best strategy so far involves a long pointed stick 2010-09-15T03:36:37 I like this contest more than tron precisely because the search space is so overwhelming 2010-09-15T03:36:55 thus there is no clear approach 2010-09-15T03:37:06 *** Palmik has quit IRC (Ping timeout: 276 seconds) 2010-09-15T03:37:18 For some reason, I haven't been able to get a bot to submit for the last 12 hours or so. I've retried several times. 2010-09-15T03:37:26 long pointy sticks have a clear and parabolic approach 2010-09-15T03:37:28 *** ShadeS has joined #aichallenge 2010-09-15T03:37:30 hello 2010-09-15T03:37:42 dufflebunk: why don't you use the tcp server? 2010-09-15T03:37:42 I downloaded the cpp starter and I'm having issues 2010-09-15T03:37:49 *** delt0r___ has joined #aichallenge 2010-09-15T03:39:17 retybok: The one at benzedrine? I can't get to it... I'm in china at the moment, so I may be blocked from it. 2010-09-15T03:39:42 *** delt0r_ has quit IRC (Ping timeout: 276 seconds) 2010-09-15T03:39:56 It may still be down 2010-09-15T03:40:03 There were some other servers I forgot the address of.. 2010-09-15T03:40:56 undefined reference to `WinMain@16' 2010-09-15T03:41:24 benzedrine is up for me. 2010-09-15T03:42:07 Probably just great firewall of china problem for me then, for benzedrine. 2010-09-15T03:42:10 benzedrine is up for me now, too. was down earlier. 2010-09-15T03:42:16 bartwe: your 7.7.1 is really bad i actually beat it this time 2010-09-15T03:42:28 oo, you're right, it is back up 2010-09-15T03:42:31 and the bot im using took 2 mins to make from stater package 2010-09-15T03:42:58 and I took a whiz in your wheaties 2010-09-15T03:43:19 thanks, that'll keep me going until the bot submission start working reliably 2010-09-15T03:43:27 Baph: thanks 2010-09-15T03:43:28 for my bot, bartwe.6 is the toughest version 2010-09-15T03:43:35 *** StanleyMT has joined #aichallenge 2010-09-15T03:43:52 apart from that one the others still kick ass 2010-09-15T03:43:55 rwa: yeah, that points to a bug in .7 it should be the same version only differing in allocation 2010-09-15T03:45:04 i should start capturing traces for replay 2010-09-15T03:45:20 make sure they are identical 2010-09-15T03:46:12 benzedrine is down again ? 2010-09-15T03:46:13 *** StanleyMT has quit IRC (Client Quit) 2010-09-15T03:46:32 *** hellman has joined #aichallenge 2010-09-15T03:47:08 bartwe_: I'm still playing a game. 2010-09-15T03:47:22 bartwe_: Maybe the web part just went south? 2010-09-15T03:47:24 me too, but can't get the website 2010-09-15T03:47:39 n player would be mega, cause bots could form coalitions 2010-09-15T03:47:54 then dog each other 2010-09-15T03:48:41 site is up for me 2010-09-15T03:49:00 @whatis tcpserver 2010-09-15T03:49:00 greghaynes: tcpserver = http://www.benzedrine.cx/planetwars/ 2010-09-15T03:49:31 antimatroid: I should try and modify tcp.c to run 3-player games and run another tcp server. 2010-09-15T03:49:58 bss03: too much effort to fix up bots during contest 2010-09-15T03:50:05 i'd be keen after, make it more than 3 player though 2010-09-15T03:50:18 like randomly generate n in [3,7] or something then run them 2010-09-15T03:50:19 antimatroid: Not sure how to Elo on 3-player matches, I could at least list recent games. 2010-09-15T03:50:34 just show the matches i suspect 2010-09-15T03:50:43 could do some ranking based on places in each match 2010-09-15T03:51:01 antimatroid: Well, I could just run them on multiple ports 9993 thorugh 9999. Players = port % 10. 2010-09-15T03:51:39 i meant fixing up my bot to handle more players :P 2010-09-15T03:52:52 *** retybok_ has joined #aichallenge 2010-09-15T03:53:42 *** _Utkarsh <_Utkarsh!~admin@117.201.86.237> has joined #aichallenge 2010-09-15T03:55:29 how do you get your bot onto the tcpserver? 2010-09-15T03:55:44 compile tcp.c, and run it 2010-09-15T03:56:14 its awesome, no waiting around for games anymore :D 2010-09-15T03:56:18 how to run? is it c? 2010-09-15T03:56:29 Yep 2010-09-15T03:56:33 Baphomet: yea --I wish more players would use thou 2010-09-15T03:56:41 Theres also a magical gui client - http://github.com/greghaynes/QTcpPlanetWars 2010-09-15T03:56:57 vizier: just compile {gcc tcp.c -o tcp} 2010-09-15T03:57:11 then run 2010-09-15T03:58:08 tcp 213.3.30.106 9999 delt0r -p 57bobbob57 runMyBot.sh 2010-09-15T03:58:22 But you would use a real password 2010-09-15T03:58:48 but --not a real one like from the web site 2010-09-15T03:58:53 its not really secure 2010-09-15T03:58:57 but no one cares 2010-09-15T03:58:58 what.... im not using a password 2010-09-15T03:59:00 is that bad? 2010-09-15T03:59:02 na 2010-09-15T03:59:18 I was testing to see if i had a new version 2010-09-15T03:59:34 I don't use that combination on the server anymore 2010-09-15T03:59:51 oo, I beat bartwe! 2010-09-15T03:59:56 w000t 2010-09-15T04:00:11 Baphomet: I'm not using a password either. I assume that means it is eay to impersonate my bot. 2010-09-15T04:01:22 delt0r: how to compile? 2010-09-15T04:01:32 gcc tcp.c 2010-09-15T04:02:04 if you are on windows then your borked. Its not like it has a real compiler ;) 2010-09-15T04:02:29 just kidding... but you probably need cygwin or something for a free compiler 2010-09-15T04:02:29 delt0r: :( yes im using windows... im to lazy to boot up linux 2010-09-15T04:03:23 tcp.c is 100% is probably just ANSI C so it should just compile on anything 2010-09-15T04:03:33 god 2010-09-15T04:03:38 good 2010-09-15T04:06:19 no perl support? 2010-09-15T04:07:26 Accoun: there is something on the forums today 2010-09-15T04:07:37 *** EvGenius has joined #aichallenge 2010-09-15T04:07:37 huh? 2010-09-15T04:07:42 at least one unoffical perl starter package 2010-09-15T04:07:43 Perl bots or perl client? 2010-09-15T04:08:21 2, i make anozer 1 2010-09-15T04:08:28 Accoun: adding new languages has a pretty high priority but getting the server stable has the highest :) 2010-09-15T04:08:53 http://ai-contest.com/forum/viewtopic.php?f=18&t=621 2010-09-15T04:09:36 translation from canadian "adding new languages has a pretty high priority but getting the server unstable has the highest" 2010-09-15T04:09:56 hmm does tcpserver support python? i have EOF error :( but playgame.jar is ok 2010-09-15T04:10:08 hellman: ?? 2010-09-15T04:10:11 The gui client does 2010-09-15T04:10:23 You should just need to chmod +x your MyBot.py 2010-09-15T04:10:48 *** caytchen has joined #aichallenge 2010-09-15T04:11:51 Wow, it seams I introdused a small bug in my bot. Which gave everything the same weighting. And it was still doing OK. I'm hoping it starts doing better now that I've fixed the bug. 2010-09-15T04:12:08 hellman: Also the worng kind of new lines can stuff the #!/bin/python up 2010-09-15T04:12:17 genericbob: are you generic on the main site? 2010-09-15T04:12:21 basically it sees only one line... which is a comment 2010-09-15T04:16:50 tr '\r\n' '\n' worked 2010-09-15T04:17:00 cool 2010-09-15T04:17:15 lots of folks are having problems with is... 2010-09-15T04:18:14 the assumption is that python shouldn't be affected by platform dependent line breaks i guess 2010-09-15T04:18:24 but the shell is 2010-09-15T04:18:53 yeah :( 2010-09-15T04:20:34 *** Pitel has joined #aichallenge 2010-09-15T04:20:52 antimatroid: I'm genericbob on the main site. No submission as of yet since my bot is in PHP. 2010-09-15T04:21:03 can someone please look what's wrong with my bot again? http://ai-contest.com/profile.php?user_id=7372 2010-09-15T04:22:38 Got enough ships yet?: http://www.benzedrine.cx/planetwars/canvas?game_id=1284536173|dhartmei|aerique 2010-09-15T04:23:48 hehe 2010-09-15T04:24:04 7.5k ships at one planet 2010-09-15T04:24:04 i shouldn't be laughing this much 2010-09-15T04:26:46 dhartmei's a-chargin' his lazah 2010-09-15T04:27:46 how to compile with cygwin? 2010-09-15T04:28:20 vizier: no idea... I guess if gcc is there.. you open a bash shell and do the gcc tcp.c thing 2010-09-15T04:28:48 why game is so long ? there's no timeout or turn limit ? tcpserver 2010-09-15T04:29:04 hellman ping time 2010-09-15T04:29:24 you can +0.5 seconds for each of my moves 2010-09-15T04:29:33 900 ms 2010-09-15T04:29:33 delt0r: i am trying... 2010-09-15T04:30:17 vizier: I can't really help with windows sorry. I don't even boot to play games. I didn't even get StarCraft 2 2010-09-15T04:30:30 delt0r: ok 2010-09-15T04:31:35 hellman: 3 minutes for a turn on tcp server 2010-09-15T04:32:06 vizier: tcp.c contains compilation line in comments 2010-09-15T04:32:19 gcc tcp.c worked for me iirc 2010-09-15T04:32:32 i understood, each turn lasts at least more than ping time, so it's because i am too far 2010-09-15T04:32:41 i tried to compile with dev c++ as a last hope before booting ubuntu :p failed 2010-09-15T04:33:08 you need at least cygwin for that to work 2010-09-15T04:33:19 tcp.c uses dup2 2010-09-15T04:33:26 i have cygwin 2010-09-15T04:33:27 not really available on windows 2010-09-15T04:33:37 gcc.... command not found 2010-09-15T04:34:08 vizier: did you open bash? 2010-09-15T04:34:25 EvGenius: cygwin... im on windows 2010-09-15T04:34:41 vizier: cygwin makes a shortcut for shell usually after installation 2010-09-15T04:35:00 *** sigh has joined #aichallenge 2010-09-15T04:35:18 vizier: you need to open this shell first 2010-09-15T04:36:27 EvGenius: i have a shell open with '~' at the top 2010-09-15T04:37:11 vizier: and this shell says gcc not found? 2010-09-15T04:37:27 vizier: did you install gcc package for cygwin? 2010-09-15T04:37:39 EvGenius: i believe so 2010-09-15T04:37:48 *** Naktibalda has joined #aichallenge 2010-09-15T04:38:45 http://ai-contest.com/starting_your_own.php I'm trying this raising value from 1 to 2 and I tried 4 just now and it's not reflecting my changes after i recompile 2010-09-15T04:40:03 then you be doing it wrong 2010-09-15T04:40:11 EvGenius: actually possibly not, where would i find it> 2010-09-15T04:41:22 vizier: I think --thought cygwin had like a update/automagic download and install everything fetrue? 2010-09-15T04:41:22 any ideas? 2010-09-15T04:41:32 I changed the 1 to a 2 and I still don't see more than 1 fleet 2010-09-15T04:41:33 so did i 2010-09-15T04:41:34 ShadeS: how are you recompiling? 2010-09-15T04:41:43 g++ mybot.cc planetwars.cc -o mybot1.exe 2010-09-15T04:41:46 well, when you install cygwin, you can select which packages are you going to install 2010-09-15T04:41:50 ShadeS: you are recompiling it wrong or executing a wrong file 2010-09-15T04:41:57 java -jar tools/PlayGame.jar maps/map6.txt 1000 1000 log.txt "java -jar 2010-09-15T04:42:04 example_bots/RandomBot.jar" MyBot1.exe | java -jar tools/ShowGame.jar 2010-09-15T04:42:11 is my batch file for it 2010-09-15T04:42:16 ShadeS: does it overwrite files? 2010-09-15T04:42:17 vizier: post somewhere the result of cygcheck -c -d 2010-09-15T04:42:29 mybot.exe was the 1 value, mybot1.exe had the value set to 2 2010-09-15T04:42:36 it shouldn't need to over write anything 2010-09-15T04:42:55 ShadeS: to be on the same side, delete the old mybot1.exe before compiling the new, or check the timestamp to know it actually has changed 2010-09-15T04:43:08 ShadeS: umm ... 2010-09-15T04:43:17 ShadeS: did you change anything in the code other than the number of fleets? 2010-09-15T04:43:28 EvGenius: no for some reason i didn't get it... 2010-09-15T04:43:37 ShadeS: are you sure you saved your files before compiling? 2010-09-15T04:43:59 just the fleet #, yes I saved 2010-09-15T04:44:06 EvGenius: not sure how to copy out of cygwin 2010-09-15T04:45:18 what's wrong? 2010-09-15T04:46:16 me? 2010-09-15T04:46:26 vizier: try clicking the small icon on the left side of the command window title. then in the menu try something like "change" => "select all" 2010-09-15T04:46:40 vizier: then select the text and press enter 2010-09-15T04:46:59 sinan: any other ideas? 2010-09-15T04:47:44 there is a menu? 2010-09-15T04:48:01 ShadeS: in my experience, when this happens it is just normally something very trivial that you forgot, for example, you might have 2 copies of the source code; you are editing one of them but compiling another one 2010-09-15T04:48:13 EvGenius: Got it 2010-09-15T04:48:20 I have one instance of codeblocks open 2010-09-15T04:48:24 ShadeS: while in the folder where you ran g++, use "less" to check the source code of mybot.cc 2010-09-15T04:48:36 ShadeS: yeah, but you are invoking g++ from cygwin, right? 2010-09-15T04:48:47 ShadeS: maybe you are editing files elsewhere :) 2010-09-15T04:49:03 ShadeS, this is a huge shot in the dark, but try adding several lines of comments to your source file and then save it again... like give it an extra hundred bytes or so. Then recompile 2010-09-15T04:49:19 EvGenius: How to send you a private message on here? 2010-09-15T04:49:21 I have some completely unfounded suspicions about code::blocks 2010-09-15T04:49:22 sinan: I'm running g++ from the command line 2010-09-15T04:49:33 EvGenius: dont want to just paste 2010-09-15T04:50:19 ShadeS: great. in the command line, write "less mybot.cc" (i think windows has the "less" command) 2010-09-15T04:50:28 vizier: depends on the type of your IRC client 2010-09-15T04:50:33 ShadeS: it should show you the content of mybot.cc inside the command prompt 2010-09-15T04:50:48 dualbot 91/100, ragebot 98/100 - code uploaded and I am off to sleep - good end to the night 2010-09-15T04:50:51 ShadeS: this way you can confirm the file you are compiling actually includes the changes :) 2010-09-15T04:50:53 vizier: you can use http://pastebin.com/ 2010-09-15T04:50:55 EvGenius: I'm using Webchat :( and i cant find a thingo for it 2010-09-15T04:50:57 ShadeS: this is a long shot, but if i was in that sitution, that's what i would do :) 2010-09-15T04:51:14 *** genericbob has quit IRC (Quit: leaving) 2010-09-15T04:51:17 Hey guys, if my bot has not participated in any competition (at ai-contest) since several hours.......could there be anything wrong with the bot itself? 2010-09-15T04:51:19 it was some bug 2010-09-15T04:51:22 http://pastebin.com/CbMMFTeT 2010-09-15T04:51:24 i closed the file in codeblocks 2010-09-15T04:51:29 Cyndre: impressive 2010-09-15T04:51:29 reopened it in codeblocks and it still said 1 2010-09-15T04:51:33 even though i saved it otherwise 2010-09-15T04:51:39 Cyndre: :) 2010-09-15T04:51:39 ah-ha, I knew it! 2010-09-15T04:51:57 ShadeS: there you have it ! ;) 2010-09-15T04:52:00 EvGenius: brb 2010-09-15T04:52:01 ty, been smashing my head against it for a while 2010-09-15T04:52:02 not yet 2010-09-15T04:52:09 reopening the file now shows >1 2010-09-15T04:52:12 why do my bot lose? http://www.benzedrine.cx/planetwars/canvas?game_id=1284539703|hellman|bartwe.7.3 i suppose it crashed, but no error messages in tcpserver 2010-09-15T04:52:13 i didn't verify it actually works yet 2010-09-15T04:52:39 vizier: I think you need to install at least these packages: gcc, binutils, w32 or win32-something 2010-09-15T04:52:41 Guys, do any of you try to run their bot against itself? I found that this is helps in finding hidden assumptions you might have made :) 2010-09-15T04:53:30 it finished in a draw 2010-09-15T04:54:01 my bot's so good, it always beats itself in an against-itself match 2010-09-15T04:54:10 *** damag has quit IRC (Ping timeout: 252 seconds) 2010-09-15T04:54:12 whiteside, +1 2010-09-15T04:54:15 :D 2010-09-15T04:54:17 *** HakanD has quit IRC (Ping timeout: 240 seconds) 2010-09-15T04:54:57 whiteside: lol 2010-09-15T04:54:59 so I think codeblocks has a bug 2010-09-15T04:55:05 I'm not going to try to recreate it 2010-09-15T04:55:07 hellman: maybe your client DCed or timed out? 2010-09-15T04:55:16 because I think it had something to do with a fresh installation of it 2010-09-15T04:55:23 eitherway I figured out what was wrong and why it's working now 2010-09-15T04:55:44 ShadeS: perfect :) 2010-09-15T04:55:48 yeah 2010-09-15T04:55:58 if it keeps borking I'll file a bug 2010-09-15T04:55:59 maybe 2010-09-15T04:56:07 I cant wait to see some games with my new bot 2010-09-15T04:56:25 *** dhaval has quit IRC (Quit: Page closed) 2010-09-15T04:56:58 ok 2010-09-15T04:57:10 now all I need to do is code in the art of war and that prob&stats book from college 2010-09-15T04:57:24 and I should be set for the top ten 2010-09-15T04:58:22 ShadeS: lol, good luck ! ;) 2010-09-15T04:58:49 whiteside: btw, i was not kidding :) if your bot only considers data on the map, you should get 100% draw. When you don't, you'll figure something might be off 2010-09-15T04:58:57 won my first one on the tcp server 2010-09-15T04:59:23 Cyndre: what's the name of your bot? 2010-09-15T04:59:40 Cyndre 2010-09-15T04:59:40 will link as soon as its up 2010-09-15T04:59:52 bartwe.7 beat me 2010-09-15T04:59:58 Cyndre: ah i see, cause i looked on the rankings page of the tcp server, didn't find it :) 2010-09-15T05:00:57 I don't see 'search by username' on their site 2010-09-15T05:01:06 http://www.benzedrine.cx/planetwars/canvas?game_id=1284540412|Cyndre|DaTwinkDaddy-s8 2010-09-15T05:01:23 sinan: stats are updated periodically 2010-09-15T05:01:40 every 7 minutes 2010-09-15T05:01:44 ahhh... I see the ol' vortex is back 2010-09-15T05:01:51 * whiteside hears his ratings swirling down the crapper 2010-09-15T05:01:52 http://www.benzedrine.cx/planetwars/getplayer?player=Cyndre 2010-09-15T05:01:54 whiteside: I don't see you at 140 any longer 2010-09-15T05:02:24 ShadeS: resubmitted code, won't be ranked until the next matchup 2010-09-15T05:02:30 oh 2010-09-15T05:02:35 *** EvGenius has left #aichallenge 2010-09-15T05:03:21 need to take more planets faster 2010-09-15T05:05:23 *** kozlovsky has joined #aichallenge 2010-09-15T05:08:21 *** vikhyat has joined #aichallenge 2010-09-15T05:11:16 Cyndre: the ultimate goal :P 2010-09-15T05:13:15 if i pre-caclulate some stuff based on the map, then identify the map and use the precalculated values, is it considered cheating? 2010-09-15T05:13:43 sinan: that is the point :P 2010-09-15T05:13:54 oh wait 2010-09-15T05:13:55 *** HakanD has joined #aichallenge 2010-09-15T05:13:58 umm no i don't think so 2010-09-15T05:14:04 antimatroid: ? is that a yes or a no? :P 2010-09-15T05:14:06 EvGenius: HOw to install those packages? 2010-09-15T05:14:09 antimatroid: ah ok :) 2010-09-15T05:14:20 that's considered fair game, but don't expect those to be the only maps used 2010-09-15T05:14:39 *** Obeleh has quit IRC (Remote host closed the connection) 2010-09-15T05:15:12 * whiteside hatches a plan to launch an anti-server, wherein more planets faster is the ultimate anti-goal, and the top rankings go to those who lose the most... The rules are that a minimum of ten ships must go out every ten turns. On the surface, it appears rather amusing... But really, it's all an attempt to divert people from the main strategy to do the exact opposite. 2010-09-15T05:15:15 antimatroid: yeah, but it can still be useful for whatever subset of the current maps show up in the final maps :) 2010-09-15T05:15:58 meh, i just assume my strategy must work for any general map 2010-09-15T05:16:03 I would have though random maps would be more fun 2010-09-15T05:16:10 thought 2010-09-15T05:16:23 delt0r___: I agree, but the organisers didn't do that last time 2010-09-15T05:16:31 at least a new set of maps including some random 2010-09-15T05:16:52 sinan: I don't think it's cheating, but note that they may change maps for the final round 2010-09-15T05:17:02 That's what they did last time at least 2010-09-15T05:17:17 I fully expect fresh random maps, myself 2010-09-15T05:17:40 Anyways a good strategy should work on all maps 2010-09-15T05:17:48 retybok: changed all the maps? 2010-09-15T05:18:06 most of them at least, I don't remember exactly 2010-09-15T05:18:16 retybok: yeah, but it might take more than 1 seconds per turn, my point is to save some time by pre-calculating some values :) still a rough idea 2010-09-15T05:18:19 What might really bork everyone is if they had very unrandom maps in the finals... like they spell words or something. That would probably throw a few bots off kilter 2010-09-15T05:18:44 whiteside: the final maps are not yet guaranteed to be symmetric? 2010-09-15T05:19:06 I think enoumous maps with say 16 players at once and over 1000 planets spanning 100 or more units would be way cool 2010-09-15T05:19:18 in fact my random map gen does just that 2010-09-15T05:19:28 as far as I know, they're not guaranteed to be anything except valid maps 2010-09-15T05:20:37 whiteside: how can you make sure a map is "fair" when it is not symmetric? 2010-09-15T05:20:59 whiteside: maybe play it twice, each time with a different starting position? 2010-09-15T05:21:01 sinan: You can't 2010-09-15T05:21:26 but if you play a lot of games... its "fair" in the totals 2010-09-15T05:21:32 I assume if that were the case, many rounds could be played 2010-09-15T05:21:33 delt0r___: and shouldn't the organizers try to make the contest fair? :) 2010-09-15T05:22:02 But a single game being unfair is not the same as the contest being unfair 2010-09-15T05:22:02 delt0r___, whiteside: yeah, i see the point. it adds some overhead though; having to play more games 2010-09-15T05:22:17 that's the beauty of this sort of contest between programs vs. real hyoonims; it's trivial to run, say, a hundred matches quickly 2010-09-15T05:22:22 delt0r___: someone should set up a server with large maps 2010-09-15T05:22:31 and how well does a bot do in cases where it is at a disadvantage? 2010-09-15T05:22:35 although more than two players would require substantial changes 2010-09-15T05:22:35 its interesting 2010-09-15T05:22:51 I think my bot would be quite good at that :) 2010-09-15T05:22:53 retybok: My code can handel it 2010-09-15T05:22:59 *** Yoshi-TS4 has quit IRC (Quit: Minus One IRC) 2010-09-15T05:23:03 I meant changes in the tcp server 2010-09-15T05:23:10 oh 2010-09-15T05:23:14 year 2010-09-15T05:23:41 whiteside: maybe at least, if they want to run non-symmetric maps at the end, they should start introducing some in the starter package. Just a suggestion :) 2010-09-15T05:23:52 perhaps i'll add a server mode to my gameplayer code 2010-09-15T05:24:34 sinan: I would relax --it will be random symmetric at worse --they won't do true random 2010-09-15T05:24:46 @tcpserver 2010-09-15T05:24:47 krakrjak: tcpserver = http://www.benzedrine.cx/planetwars/ 2010-09-15T05:24:57 for n players I would use n fold symmetry 2010-09-15T05:27:50 *** num1_ has quit IRC (Ping timeout: 245 seconds) 2010-09-15T05:29:32 *** xha has joined #aichallenge 2010-09-15T05:30:25 man 2010-09-15T05:30:32 compiling time is great for cigarette rolling 2010-09-15T05:31:28 I don't think i have had to wait for code to compile for years 2010-09-15T05:31:36 no incremental compilation 2010-09-15T05:31:41 this is a some what modern machine 2010-09-15T05:31:44 it takes a few seconds 2010-09-15T05:31:55 watching this thing play out will definitly take more time than to roll a cigarette 2010-09-15T05:33:49 for now my multiple fleets seem to head towards the same planet... 2010-09-15T05:34:29 not so bad as for the first change :) 2010-09-15T05:34:41 *** EvGenius has joined #aichallenge 2010-09-15T05:35:25 Does anybody know if the euclidian distance is equal to the number of turns required to send ships, or is there another formula? 2010-09-15T05:35:43 (send ships to a planet if that wasn't obvious) 2010-09-15T05:35:56 ceil(euclidean distance) 2010-09-15T05:36:04 Thanks. 2010-09-15T05:36:35 thanks for reminding me about my analysis assignment :( 2010-09-15T05:37:20 * kjeldahl grins 2010-09-15T05:44:08 i'm looking at this code on the simple strat 2010-09-15T05:44:36 pw.Production is giving me trouble in c++ 2010-09-15T05:47:38 it's a simple loop 2010-09-15T05:47:50 isn't it? 2010-09-15T05:48:39 *** dufflebunk has quit IRC (Ping timeout: 252 seconds) 2010-09-15T05:49:12 i'd like to help get the ruby starter package out there, what can i do? 2010-09-15T05:49:17 the syntax was for java 2010-09-15T05:49:28 i changed boolean to bool 2010-09-15T05:49:36 but I don't know if i need to changeanything else 2010-09-15T05:50:15 oh, why do you use java starter package in c++? 2010-09-15T05:50:29 *** retybok has quit IRC (Quit: leaving) 2010-09-15T05:51:13 the http://ai-contest.com/simple_strategy_guide.php shows it in java 2010-09-15T05:52:49 sinan: make one and put it on forum (: 2010-09-15T05:52:57 *** caytchen has quit IRC (Ping timeout: 240 seconds) 2010-09-15T05:53:04 or ask in forum if somebody else is working on it 2010-09-15T05:53:11 HakanD: i mean, if a lot of work has already been done there, it would be awaste of time 2010-09-15T05:53:15 HakanD: cool, will do that 2010-09-15T05:53:36 HakanD: just thought i might get a faster response here, hehe :) 2010-09-15T05:53:59 u ll get probably, just not now (: 2010-09-15T05:54:12 HakanD: found a thread related to this: http://ai-contest.com/forum/viewtopic.php?f=18&t=401 2010-09-15T05:54:15 thinking about time zones 2010-09-15T05:54:23 guys, how often organizations are updated? :/ 2010-09-15T05:54:41 it's a manual task :) 2010-09-15T05:57:01 HakanD: turns out someone already built a starter package, but the server environment doesn't support ruby yet :) 2010-09-15T05:57:14 ShadeS: it looks like you have issue with understanding basics of c++ - flow control, object method call 2010-09-15T05:57:47 I'm a bit rusty on my c++ 2010-09-15T05:57:57 sinan: yep, adding new languages are near top of the list, but it's never done :( 2010-09-15T05:58:05 ShadeS: choose another language 2010-09-15T05:58:16 Naktibalda: can i help to accelerate the process? 2010-09-15T05:58:26 in codeblocks pw. brings up a list of items but I don't see production to be one of them 2010-09-15T05:58:36 find someone responsible and kick his ass, physically 2010-09-15T06:00:08 http://code.google.com/p/ai-contest/issues/detail?id=55 2010-09-15T06:00:38 see, it was created 9 days ago 2010-09-15T06:01:20 Naktibalda: lol, yeah, but last comment was yesterday, I imagine you guys are quite busy ;) 2010-09-15T06:01:32 not me 2010-09-15T06:01:37 Naktibalda: as long as it will happen sometime, i can code in ruby and use the tcp server for testing :) 2010-09-15T06:01:41 what timezone is the ai-contest server in? 2010-09-15T06:01:48 *** boegel has joined #aichallenge 2010-09-15T06:01:54 hiya y'all 2010-09-15T06:02:08 is it specified somewhere what is exactly needed in the ZIP file? 2010-09-15T06:02:11 Naktibalda: I think I'll stick with c++ 2010-09-15T06:02:16 sinan: I hope so too. I'm waiting for PHP support 2010-09-15T06:02:20 as I know even less about java and python 2010-09-15T06:02:27 of course, the bot's code is, but also additional files to get it to compile and such? 2010-09-15T06:03:31 boegel: whatever is needed to compile your files 2010-09-15T06:03:40 boegel: what additional files do you need? 2010-09-15T06:03:51 boegel: that's normally your code, and things like the panetwars class 2010-09-15T06:04:17 is here a some naming convention? how do they know what files to compile? 2010-09-15T06:04:31 yes, there is 2010-09-15T06:05:24 Naktibalda: they compile all files that are ".", then link all of them together 2010-09-15T06:06:09 Naktibalda, boegel: check "Do we have to submit PlanetWars.java/ PlanetWars.cc" in http://ai-contest.com/faq.php 2010-09-15T06:06:12 Your main code file must be called MyBot.ext, where ext is java, cc, or whatever the standard extension for code files in your language is. Remember to include all the code files that your entry needs. 2010-09-15T06:07:33 sinan: k, great, thx 2010-09-15T06:07:38 EvGenius: no idea :) 2010-09-15T06:09:12 *** awwaiid has quit IRC (Ping timeout: 240 seconds) 2010-09-15T06:10:08 *** fx4m has joined #aichallenge 2010-09-15T06:11:14 *** awwaiid has joined #aichallenge 2010-09-15T06:11:51 *** xha has quit IRC (Quit: Odcházím) 2010-09-15T06:13:55 I'm trying to find information about this pw.Production.. 2010-09-15T06:14:29 it's in PlanetWars.cc file 2010-09-15T06:15:16 i'm searching for the string production and it's not finding it in that file 2010-09-15T06:15:56 do you see it in there? 2010-09-15T06:16:03 anyone know DualBot algoritm? 2010-09-15T06:16:50 Accoun: example_bots/DualBot.java 2010-09-15T06:18:22 planetwars.java does but not planetwars.cc... 2010-09-15T06:18:34 correct 2010-09-15T06:18:39 it was never implemented for c++ 2010-09-15T06:19:13 right, Baphomet can help about c++ 2010-09-15T06:19:23 oh 2010-09-15T06:19:52 but the function in java is pw.Production(1) for player 1 and will return total population of all planets 2010-09-15T06:20:02 owned by that player 2010-09-15T06:20:32 i tried to copy java and covert the code to c++ and add it into the headers but it was all like: no you cant do that/ your doing it wrong 2010-09-15T06:20:32 when I change number of fleets and I run it initially there is a stream of ships heading to the same planet. I can change this by doing? 2010-09-15T06:21:26 well thats because in the single turn, it keeps the same dest and if you have set it to allow multiple active fleets it will send however mant is possible 2010-09-15T06:22:06 oh, is a streaming feature built-in in C++ package? 2010-09-15T06:22:30 streaming? 2010-09-15T06:22:35 where it finds the strongest planet.. if I have 4 fleets out for my first turn, can it be possible to find the next strongest planet after the first to have them all go to separate locations? 2010-09-15T06:22:55 yes that is possible but will require some more coding 2010-09-15T06:23:00 in a single turn 2010-09-15T06:23:03 yea 2010-09-15T06:23:26 just reuse the code to find dest and modify it to how you want inside of a loop 2010-09-15T06:23:58 ok i'm going to look at that when i wake up then 2010-09-15T06:24:20 I've gotten farther than I anticipated tonight 2010-09-15T06:25:41 OMG, my bot won against bartwe.7.8.1 once, in 300 turns 2010-09-15T06:26:11 very nice 2010-09-15T06:27:53 whois that 2010-09-15T06:28:09 *** HakanD has quit IRC (Read error: Connection reset by peer) 2010-09-15T06:28:20 the latest version of bartwe bot :) 2010-09-15T06:28:23 bartwe's bot, quite a powerfull one that is dificult to beat 2010-09-15T06:29:03 yeah, and mine is a rather weak one 2010-09-15T06:29:07 link of it in action? 2010-09-15T06:29:23 *** zaph has joined #aichallenge 2010-09-15T06:29:26 http://www.benzedrine.cx/planetwars/canvas?game_id=1284544369|Naktibalda-002|bartwe.7.8.1 2010-09-15T06:32:36 *** zaph has quit IRC (Client Quit) 2010-09-15T06:33:03 i kill bartwe earlier today too for the first time http://www.benzedrine.cx/planetwars/canvas?game_id=1284535765|Baphomet-ai-v5|bartwe.7.7.1 2010-09-15T06:33:21 stealing main planet is a hell cheap strat tho hehehehe 2010-09-15T06:34:47 yep, I think it's a good idea to atack his planet on the first turn, you ether gain a lot or it's a draw in ship count 2010-09-15T06:35:26 unless enemy is far away and around starting planet are a lot of cheap planets for expansion 2010-09-15T06:39:12 it would be possible to get a symetric game if both bots use this tactic all game :) 2010-09-15T06:44:46 *** HakanD has joined #aichallenge 2010-09-15T06:45:15 *** pvdbrand has joined #aichallenge 2010-09-15T06:45:15 *** anoek has quit IRC (Ping timeout: 264 seconds) 2010-09-15T06:45:23 *** mightybyte has left #aichallenge 2010-09-15T06:45:43 *** anoek has joined #aichallenge 2010-09-15T06:46:03 Blah. That took longer than it should have. 2010-09-15T06:46:16 Probably because I keep getting distracted by pretty visualizations. 2010-09-15T06:47:09 Still, finally got the simulator part of my bot written. Now, I just need to figure out how best to modify by planet value function. 2010-09-15T06:47:19 *** justin_pdx has joined #aichallenge 2010-09-15T06:47:28 Hopefully, I'll be climbing the ranking again tomorrow. 2010-09-15T06:48:05 *** justin_pdx has quit IRC (Client Quit) 2010-09-15T06:51:56 what does the "oppo." column mean on the benzedrine rankings page 2010-09-15T06:52:19 Is it possible that submitting bots is not working at all at the moment? I know the email notification doesn't work and that it's being worked on, but even submitting my previous submission didn't work anymore; it doesn't show up in my list of submissions even after a couple of hours... 2010-09-15T06:54:13 *** Bobng has joined #aichallenge 2010-09-15T07:00:23 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-09-15T07:01:31 *** kozlovsky_ has joined #aichallenge 2010-09-15T07:02:14 aerique: I have cleaned up the cl starter kit. 2010-09-15T07:02:35 hmm http://ai-contest.com/visualizer.php?game_id=4473742 gg 2010-09-15T07:02:41 *** iFire has joined #aichallenge 2010-09-15T07:02:46 *** bss03 is now known as bss03-AFK 2010-09-15T07:02:59 no specials, socket-closes, asdf, etc 2010-09-15T07:04:12 *** kozlovsky has quit IRC (Ping timeout: 252 seconds) 2010-09-15T07:04:46 mega1: cool, posted it to the thread yet? 2010-09-15T07:05:14 not yet, I'm finishing up the readme 2010-09-15T07:05:22 mega1: also realize that you've got a pretty limited environment on the game servers (not sure what you're doing with asdf) 2010-09-15T07:06:06 can someone explain how the ranking system works? 2010-09-15T07:06:22 if your bot if found to be good, is it being played against other strong bots? 2010-09-15T07:06:24 mega1: include the updated tools with it: http://ai-contest.com/forum/viewtopic.php?f=18&t=534 2010-09-15T07:06:32 *** whoaski has joined #aichallenge 2010-09-15T07:06:36 boegel: yes 2010-09-15T07:06:38 aerique: I have done that. 2010-09-15T07:06:40 and that way, do you get ranked higher if you winagainst those strong bots? 2010-09-15T07:07:21 aerique: as to asdf, I split MyBot.lisp into several files, factored the logging out to a separate system that's used by the proxy-bot system 2010-09-15T07:07:29 *** eburnette has joined #aichallenge 2010-09-15T07:07:41 boegel: yes, it uses this: http://remi.coulom.free.fr/Bayesian-Elo/ 2010-09-15T07:07:42 EvGenius: You here? 2010-09-15T07:08:09 vizier: yes 2010-09-15T07:08:09 I'm not sure if asdf can be made to work on the server even though I added split-sequence and parse-number along with their asdf files. 2010-09-15T07:08:12 *** whoaski has left #aichallenge 2010-09-15T07:08:25 mega1: sounds great. are you playing here yet?: http://www.benzedrine.cx/planetwars/ 2010-09-15T07:08:38 EvGenius: ok i finally got gcc but where should i put the tcp file? 2010-09-15T07:08:56 vizier: you mean tcp.c? 2010-09-15T07:09:01 aerique: no, not yet. I'll go by the nick bocsimacko again. 2010-09-15T07:09:18 EvGenius: Yes 2010-09-15T07:09:30 vizier: anywhere. the only thing is to run compiling command line from the same directory 2010-09-15T07:10:06 EvGenius: ahh... sry had a temporary mind blank 2010-09-15T07:10:51 aerique: do we know the sbcl version on the server? 2010-09-15T07:12:00 EvGenius: ... now it says i cant run it becuase i dont have cygwin.dll 2010-09-15T07:12:15 *** Yoshi-TS4 has joined #aichallenge 2010-09-15T07:12:38 *** kozlovsky_ has quit IRC (Ping timeout: 252 seconds) 2010-09-15T07:12:38 mega1: not yet, probably 1.0.29 since that comes with the OS running on the server. i tried for 1.0.40 but they're really short on time 2010-09-15T07:12:51 *** kozlovsky has joined #aichallenge 2010-09-15T07:13:12 vizier: so it compiled and linked fine, but won't run? 2010-09-15T07:13:30 EvGenius: linked? yes it made a file but no it wont run 2010-09-15T07:14:29 so get cygwin.dll 2010-09-15T07:14:39 vizier: it's weird that it can't find cygwin.dll 2010-09-15T07:15:07 EvGenius: its cygwin1.dll 2010-09-15T07:16:15 vizier: try to find this file somewhere in cygwin directory tree and then copy this file into the directory in which tcp.exe resides 2010-09-15T07:16:42 *** vikhyat has quit IRC (Quit: Leaving) 2010-09-15T07:16:49 *** EvGenius has left #aichallenge 2010-09-15T07:17:19 I dont see why anyone would need to use cygwin for this? 2010-09-15T07:17:32 vizier: is cygwin.dll in your path when you execute it. 2010-09-15T07:18:17 EvGenius: i dont have it anywhere on my pc apperntly 2010-09-15T07:18:31 vizier: did you install cygwin? 2010-09-15T07:18:32 apperently its not on my computer... 2010-09-15T07:18:34 i did 2010-09-15T07:18:42 odinsbane: i did 2010-09-15T07:18:50 vizier: then you have it. How are you trying to run it? 2010-09-15T07:19:31 odinsbane: just through the exe... i did a search of my pc... no file to be found 2010-09-15T07:19:34 *** EvGenius has joined #aichallenge 2010-09-15T07:19:42 so dont run cygwin? 2010-09-15T07:20:13 everything works in windows as well as linux 2010-09-15T07:20:14 im running tcp.c 2010-09-15T07:20:15 vizier: by double clicking on it? or from a bash shell or from a cmd prompt. 2010-09-15T07:20:27 vizier im using tcp.c in windows with no issues 2010-09-15T07:20:40 odinsbane: i just double clicked on it 2010-09-15T07:21:11 vizier: so, did you found cygwin1.dll? 2010-09-15T07:21:12 vizier: where did you install cygwin? ie C:\cygwin\bin\cygwin1.dll should exist. 2010-09-15T07:21:58 wait it does exist... 2010-09-15T07:22:07 how should i be running tcp then? 2010-09-15T07:22:14 vizier: try running it from a bash prompt 2010-09-15T07:22:14 EvGenius: i found it 2010-09-15T07:22:15 vizier: check pm 2010-09-15T07:24:06 What's with all the sha16x bots? 2010-09-15T07:24:19 vizier: copy this dll into the directory with tcp.exe 2010-09-15T07:24:42 vizier: so it's guaranteed it'll find it 2010-09-15T07:25:10 EvGenius: He has the dll already, its probably better if he doesn't mix them. 2010-09-15T07:25:43 odinsbane: but for some reason when he runs tcp.exe it says it can't find dll 2010-09-15T07:25:52 EvGenius: plus he is going to need to run it with cmd line parameters its probably best if he just goes to an xterm and runs the program. vizier 2010-09-15T07:26:28 EvGenius: because he is double clicking on it. Your right he could copy the dll to the directory to be able to double click on it, but then how does he pass the cmd line options? 2010-09-15T07:28:35 odinsbane: sure, he'll need command line anyway 2010-09-15T07:28:58 im using command line anyway now 2010-09-15T07:30:18 for anyone else that wants the compiled tcp.c in .class (works with java in windows) form http://baphomet.dontexist.com/tcp.class 2010-09-15T07:31:33 Baphomet: somebody wrote a java version? 2010-09-15T07:32:10 ugh 2010-09-15T07:32:15 my thing crashed 2010-09-15T07:32:18 *** pvdbrand has quit IRC (Quit: Leaving) 2010-09-15T07:32:35 ShadeS: didn't you go to sleep a hour ago? 2010-09-15T07:32:47 lol 2010-09-15T07:32:54 i tried to 2010-09-15T07:33:04 and i just was tossing and turning 2010-09-15T07:33:05 but the programming told you to keep going? 2010-09-15T07:33:22 i figured if i'm going to toss and turn i might as well craack out some code 2010-09-15T07:33:22 std::cout << "wake up......."; 2010-09-15T07:33:41 how can I debug this 2010-09-15T07:34:53 *** ath888 has joined #aichallenge 2010-09-15T07:38:42 would anyone know my bot which i know works suddenly develops an unkown error when running on the tcp 2010-09-15T07:39:22 add traces 2010-09-15T07:39:47 on the tcp? 2010-09-15T07:39:50 *** jmreardon has joined #aichallenge 2010-09-15T07:40:03 just log everything 2010-09-15T07:40:03 in your bot 2010-09-15T07:40:09 it shouldn't error 2010-09-15T07:40:16 but it does 2010-09-15T07:40:20 i have tested it 1000 times and it works everyimt 2010-09-15T07:40:26 i have tested it 1000 times and it works everytime 2010-09-15T07:40:29 lol 2010-09-15T07:40:37 and... it does the same thing with 2 different bots of mine 2010-09-15T07:40:41 both of which work 2010-09-15T07:40:42 because your tests were flawed 2010-09-15T07:40:56 or... it something with how im doing it 2010-09-15T07:41:27 what error do you have? exception? timeout? other? 2010-09-15T07:41:33 unknown 2010-09-15T07:41:45 unknown exception 2010-09-15T07:42:02 which language ? 2010-09-15T07:42:13 python 2010-09-15T07:42:24 unknown exception? 2010-09-15T07:42:25 can't you just catch it and save it in file ? 2010-09-15T07:42:32 i can i guess 2010-09-15T07:42:33 what do you mean unknwon exception. 2010-09-15T07:42:35 Bobng: unknown to him :) 2010-09-15T07:42:38 does it start working or does it crash instantly as soon as you connect to another user? 2010-09-15T07:42:43 oh. thats simple. 2010-09-15T07:42:48 wrap your main turn function 2010-09-15T07:42:50 in a try/except 2010-09-15T07:42:54 crashes instantly 2010-09-15T07:43:06 vizier: how are you running it? 2010-09-15T07:43:11 if it is same than c# client you just have to add code to handle the exception 2010-09-15T07:43:21 yeah 2010-09-15T07:43:23 odinsbane: through bash 2010-09-15T07:43:25 instead of the stupid default // Owned "handler" 2010-09-15T07:43:28 but if it crashes instantly 2010-09-15T07:43:40 yeah... it shouldn't 2010-09-15T07:43:41 just add some logging code at the top 2010-09-15T07:43:46 vizier: is the file executeable? Do you have the #!/usr/bin/env python ? 2010-09-15T07:43:52 from twisted.python import log 2010-09-15T07:44:04 then setup a logfile and put lots of "log.msg" calls everywhere 2010-09-15T07:44:23 from the code that shows me the weakest enemy or neutral planet.. how could I find the next weakest? 2010-09-15T07:44:25 oqindsbane: yes it is 2010-09-15T07:44:40 where is the information for finding out how many planets are in the game? 2010-09-15T07:44:41 hmm... 2010-09-15T07:44:58 ShadeS: do a list of planet and sort them according to their strength take second in list 2010-09-15T07:44:58 yeah but crashing instantly... there has got to be a problem somewhere.... 2010-09-15T07:44:59 vizier: it is also running cygwins python now and not your windows python. 2010-09-15T07:45:01 *** boegel has left #aichallenge ("Leaving") 2010-09-15T07:45:23 odinsbane: ahh... no it isn't... it was using windows python 2010-09-15T07:45:35 ShadeS: PlanetWars.Planets().Count 2010-09-15T07:45:52 hmm i think there is a PlanetCount() also 2010-09-15T07:46:07 odinsbane: where is cygwins python? can i just call python? 2010-09-15T07:46:22 vizier: probably, try which python and it should show you which one your using. 2010-09-15T07:46:30 edcba: for c++? 2010-09-15T07:46:45 edcba: ah I see 2010-09-15T07:46:45 i guess it's same everywhere 2010-09-15T07:47:45 odinsbane: still errors with cygwi's python 2010-09-15T07:48:28 vizier: can you get the stack trace, ie just run python "MyBot.py" 2010-09-15T07:49:06 *** Caffeinated has quit IRC (Ping timeout: 245 seconds) 2010-09-15T07:49:42 fixed it 2010-09-15T07:50:05 how long before the game is visible 2010-09-15T07:50:22 *** caytchen has joined #aichallenge 2010-09-15T07:50:50 how long does a game take? 2010-09-15T07:50:57 oh god, long revisions are long 2010-09-15T07:52:01 vizier: over tcp? up to 7 minutes 2010-09-15T07:52:37 usualyl less 2010-09-15T07:53:01 well, 3.5 should be close to average 2010-09-15T07:53:31 i got it now 2010-09-15T07:55:26 *** _Utkarsh <_Utkarsh!~admin@117.201.86.237> has quit IRC (Ping timeout: 265 seconds) 2010-09-15T07:55:55 vizier: what did you have to fix? 2010-09-15T07:56:35 *** tobiassjosten has joined #aichallenge 2010-09-15T07:58:43 odinsbane: i forgot to copy planet wars and that stuff 2010-09-15T08:00:56 *** sinan has quit IRC (Ping timeout: 272 seconds) 2010-09-15T08:01:01 *** Utkarsh has joined #aichallenge 2010-09-15T08:04:40 *** fucawb has quit IRC (Remote host closed the connection) 2010-09-15T08:05:57 ugh how can I debug this? 2010-09-15T08:06:10 *** otherAntimatroid has left #aichallenge 2010-09-15T08:07:45 *** EvGenius has left #aichallenge 2010-09-15T08:08:46 it's crashing and I think it's because it's refering a planetID that isn't there 2010-09-15T08:10:12 *** ath888 has quit IRC (Ping timeout: 252 seconds) 2010-09-15T08:10:41 *** jmreardon has quit IRC (Quit: jmreardon) 2010-09-15T08:13:53 *** Appleman1234 has quit IRC (Ping timeout: 240 seconds) 2010-09-15T08:14:40 *** braddunbar has quit IRC (Ping timeout: 276 seconds) 2010-09-15T08:16:11 *** amstan has joined #aichallenge 2010-09-15T08:16:11 *** ChanServ sets mode: +o amstan 2010-09-15T08:16:32 log your orders to file 2010-09-15T08:16:58 yesterday my bot crashed because it sent a negative number of ships :) 2010-09-15T08:19:47 *** Queue29 has quit IRC (Remote host closed the connection) 2010-09-15T08:20:08 how do I do this? 2010-09-15T08:20:23 my log file just says initializing 2010-09-15T08:20:28 and nothing else in it 2010-09-15T08:21:42 that's engine log file 2010-09-15T08:21:44 ShadeS: from twisted.python import log 2010-09-15T08:21:47 just write your commands to file 2010-09-15T08:21:49 then do 2010-09-15T08:21:50 log.msg 2010-09-15T08:21:57 *** vikhyat has joined #aichallenge 2010-09-15T08:22:17 after doing 2010-09-15T08:22:19 log.startLogging(open('/var/log/foo.log', 'w')) 2010-09-15T08:22:35 thats python ofc 2010-09-15T08:22:39 *** PlanetWarrior has quit IRC (Quit: Page closed) 2010-09-15T08:22:49 i'm doing c++ 2010-09-15T08:22:52 not pyhthon 2010-09-15T08:23:20 *** oh_cripes has joined #aichallenge 2010-09-15T08:23:47 *** virtualmikeai has joined #aichallenge 2010-09-15T08:24:00 sux2beu then 2010-09-15T08:24:14 wtf 2010-09-15T08:28:16 *** jmreardon has joined #aichallenge 2010-09-15T08:30:40 ShadeS, you can use this http://codepad.org/t3BoIdOb 2010-09-15T08:31:20 *** ermau has joined #aichallenge 2010-09-15T08:32:06 *** ermau is now known as ermau|work 2010-09-15T08:34:43 i put that in mybot.cc? 2010-09-15T08:38:01 http://i.imgur.com/jacoj.jpg 2010-09-15T08:38:15 this sums up what you need to do 2010-09-15T08:38:56 *** vikhyat has quit IRC (Ping timeout: 272 seconds) 2010-09-15T08:41:03 *** virtualmikeai has quit IRC (Quit: Page closed) 2010-09-15T08:41:42 *** Palmik__ has joined #aichallenge 2010-09-15T08:41:53 *** ath888 has joined #aichallenge 2010-09-15T08:42:26 *** Palmik_ has quit IRC (Ping timeout: 245 seconds) 2010-09-15T08:43:25 *** Palmik__ is now known as Palmik 2010-09-15T08:43:55 Btw... what happened to the articled about technical issues on the main page? :) 2010-09-15T08:44:26 lol, I just saw that too 2010-09-15T08:44:41 I guess we've not had many for the last two days. 2010-09-15T08:45:14 Well, website still has tons of em (: 2010-09-15T08:47:12 *** mightybyte has joined #aichallenge 2010-09-15T08:47:36 *** eburnette has quit IRC (Ping timeout: 252 seconds) 2010-09-15T08:50:52 What article? 2010-09-15T08:51:16 *** vikhyat has joined #aichallenge 2010-09-15T08:52:04 *** s_denchev has joined #aichallenge 2010-09-15T08:53:03 *** Warlam has joined #aichallenge 2010-09-15T09:02:36 *** kozlovsky has quit IRC (Quit: Page closed) 2010-09-15T09:03:41 *** HakanD has quit IRC (Ping timeout: 245 seconds) 2010-09-15T09:04:52 *** boegel has joined #aichallenge 2010-09-15T09:05:06 can anyone explain to me why red is losing here? I must be missing something 2010-09-15T09:05:07 http://ai-contest.com/visualizer.php?game_id=4476550 2010-09-15T09:06:08 *** bss03-AFK is now known as bss03 2010-09-15T09:06:12 Crash / Timeout? 2010-09-15T09:06:30 bss03: right, forgot about that, that's probably it 2010-09-15T09:08:14 this is funny, they keep fighting over the same planet :) 2010-09-15T09:08:14 http://ai-contest.com/visualizer.php?game_id=4480015 2010-09-15T09:08:24 *** awwaiid has quit IRC (Quit: leaving) 2010-09-15T09:08:31 *** braddunbar has joined #aichallenge 2010-09-15T09:17:33 stupid bots :) 2010-09-15T09:17:35 *** coventry has joined #aichallenge 2010-09-15T09:19:27 what's the latest on address space / memory limits? 2010-09-15T09:20:02 There is a legend at the bottom of http://www.benzedrine.cx/planetwars/getplayer?player=vortex.12 indicating that games which ended due to a timeout or disconnect will be annotated as such. However, the game http://www.benzedrine.cx/planetwars/canvas?game_id=1284554534|smr2-18-8|vortex.12 appears to have gone to smr2-18-8, even though vortex.12 is clearly winning on the board. Does that mean that vortex.12 made an illegal move? Would 2010-09-15T09:20:02 it be possible to include an annotation for that in the server? 2010-09-15T09:20:39 whwre is source of PlayGame.jar ? 2010-09-15T09:21:07 I haven't seen a game marked with T or D yet 2010-09-15T09:21:52 Huh. I maybe that functionality hasn't been implemented, yet. Would be nice to have. 2010-09-15T09:23:06 Accoun: looking that up for you 2010-09-15T09:23:30 Accoun: http://code.google.com/p/ai-contest/source/browse/#svn/trunk/planet_wars/viz 2010-09-15T09:24:38 *** sigh has quit IRC (Remote host closed the connection) 2010-09-15T09:26:19 yeah, you can thank me for knocking java off top place :D 2010-09-15T09:28:15 antimatroid: you're dhartmei? 2010-09-15T09:28:32 * boegel is planning to overthrow everyone else using Haskell, beat that 2010-09-15T09:28:43 nope, youwillfail 2010-09-15T09:28:51 oh noes I wonts 2010-09-15T09:29:02 you will once i improve it more 2010-09-15T09:29:17 I don't have much experience, but I'll have lot's of free time the coming week and a half :D 2010-09-15T09:29:25 but i killed tracer which i suspect pulled him under first 2010-09-15T09:30:24 antimatroid: so, you're dhartmei then? 2010-09-15T09:35:06 *** O1athe is now known as Olathe 2010-09-15T09:37:22 dhartmei is dhartmei 2010-09-15T09:38:34 *** Warlam has quit IRC (Quit: Page closed) 2010-09-15T09:39:33 mega1: there are no limits now 2010-09-15T09:39:44 mega1: other then the fact the current server only has 2gb of ram 2010-09-15T09:39:51 so that could count as a limit 2010-09-15T09:40:00 I remember that was what you said yesterday. 2010-09-15T09:40:50 dstufft: I'm finishing up the common lisp starter package and it's hard to decide if there should be limit on the heap. 2010-09-15T09:41:17 on x86-64 sbcl defaults to allocate 8G of _address space_ 2010-09-15T09:41:29 Naktibalda: I was just wondering why antimatroid stated he kicked Java off first place 2010-09-15T09:41:33 of course, it's up to the bot how much it uses of that. 2010-09-15T09:41:39 mega1: i don't know a lot about lisp but we won't be limiting AS again 2010-09-15T09:41:49 if the most we'll do is limit RSS or whatever 2010-09-15T09:42:06 boegel: because when tracer had a loss, he was no longer number 1, he was not below who he lost to. 2010-09-15T09:42:13 dstufft: that's also important to know 2010-09-15T09:42:16 *** Baphomet has quit IRC (Ping timeout: 264 seconds) 2010-09-15T09:42:31 dstufft: how many games will be run in parallel on the same box? 2010-09-15T09:43:09 mega1: atm theres no plans to limit memory, and atm to my knowledge we are only running 1 game at a time 2010-09-15T09:43:47 so if my opponent hogs the memory I'm at the mercy of the kernel? 2010-09-15T09:44:05 how much memory could one of these programs require? 2010-09-15T09:44:44 shouldn't a bot be disqualified for such behaviour if it's intentional 2010-09-15T09:45:02 I don't care if it's intentional. 2010-09-15T09:45:18 mega1: currently, i suppose we could limit RSS, if its intentional it would be DQ'd though i believe, since disruptions are against the rules 2010-09-15T09:45:28 i just woke up so bear with me :p 2010-09-15T09:46:10 well, I don't think that any sane bot could use >1G of memory 2010-09-15T09:46:17 ok, I'll put a 1G limit on the heap for now. It seems that no should go above that. 2010-09-15T09:46:33 *no bot 2010-09-15T09:46:52 *** Baphomet has joined #aichallenge 2010-09-15T09:46:57 mega1: i'll bring it up on the mailing list and see what folks think, or make a post on the forum 2010-09-15T09:47:10 for technical reasons, I can easily waste as much memory as available 2010-09-15T09:47:28 i don't know all the languages, i didn't realize java etc addressed like 700mb+ of shit in even an empty program 2010-09-15T09:47:36 so when I limited AS i killed javabots 2010-09-15T09:47:58 so i don't want to limit ram again without giving people a chance to look at the potetional limits 2010-09-15T09:48:02 lol 2010-09-15T09:51:02 *** vizier has quit IRC (Ping timeout: 252 seconds) 2010-09-15T09:51:49 and jesus 8g in the AS? wtf for >.> 2010-09-15T09:55:22 *** s_denchev has quit IRC (Quit: Page closed) 2010-09-15T09:56:22 bleh 2010-09-15T09:56:24 mega1: i think during the Tron contest they gave a --dynamic-space-size argument of 1G (or was it 512MB) to SBCL during compilation of the bot 2010-09-15T09:56:26 compiling is broken again 2010-09-15T09:56:48 aerique: I set it to 768M for the time being. 2010-09-15T09:57:45 mega1: i am not certain however if that argument to sbcl during compilation carries over to the resulting binary 2010-09-15T09:58:17 aerique: hmm, I have to check. There might have been some changes related to that. 2010-09-15T09:59:17 mega1: yeah, i thought i read something to that extend on the sbcl mailinglist. it might even be intentional 2010-09-15T09:59:42 dstufft: That java stuff is all shared lib and stuff... its not "real" usage in the normal sense... but if its running in a sand box 2010-09-15T09:59:55 delt0r___: yea i know 2010-09-15T10:00:08 delt0r___: i just didnt realize Java addressed that much 2010-09-15T10:00:19 so when i limited address space to 196M it killed them :p 2010-09-15T10:00:26 but yea.. for light weight stuff the jvm has always been on the heavy side 2010-09-15T10:02:27 i swear compile_daemon and compile_anything is terrible 2010-09-15T10:02:48 841 items in the backlog, but we are compiling again 2010-09-15T10:03:41 *** vikhyat has quit IRC (Ping timeout: 245 seconds) 2010-09-15T10:06:09 *** ath888 has quit IRC (Quit: Page closed) 2010-09-15T10:10:36 *** hornairs has quit IRC (Quit: hornairs) 2010-09-15T10:15:16 *** sinan has joined #aichallenge 2010-09-15T10:20:00 aerique: --save-runtime-options is since 1.0.31 2010-09-15T10:21:19 I mean :save-r-o 2010-09-15T10:25:24 ya'll miss me? 2010-09-15T10:27:07 *** Zeiris has quit IRC (Read error: Connection reset by peer) 2010-09-15T10:28:31 *** smhanov__ has joined #aichallenge 2010-09-15T10:29:39 *** kozlovsky has joined #aichallenge 2010-09-15T10:31:55 *** smhanov_ has quit IRC (Ping timeout: 252 seconds) 2010-09-15T10:31:59 hmm, my bot played against 7 different versions of smr2 in the last 30 minutes 2010-09-15T10:32:32 *** rogue780 has joined #aichallenge 2010-09-15T10:34:28 *** xtile has joined #aichallenge 2010-09-15T10:35:09 contestbot: rankings 2010-09-15T10:35:10 xtile: Top 10 players: dhartmei(3132), liehann(3088), Tracer(3082), _Shurf(3043), deepblue(2962), Macuyiko(2953), adumlah(2928), pdbpdb80(2913), Loks1982(2906), youwillfail(2892) 2010-09-15T10:35:34 contestbot: rankings 100 2010-09-15T10:35:34 xtile: Error: Cannot do that many players. 2010-09-15T10:35:39 Naktibalda: same here :-| 2010-09-15T10:35:39 contestbot: rankings 140 2010-09-15T10:35:40 xtile: Error: Cannot do that many players. 2010-09-15T10:35:45 contestbot: rankings 40 2010-09-15T10:35:46 xtile: Top 40 players: dhartmei(3132), liehann(3088), Tracer(3082), _Shurf(3043), deepblue(2962), Macuyiko(2953), adumlah(2928), pdbpdb80(2913), Loks1982(2906), youwillfail(2892), esrauch(2848), a1k0n(2846), damien(2841), Jerome(2839), Molehill(2833), dabino(2830), bartwe(2826), GreenTea(2816), r.nitsch(2782), sighbot(2773), Dungeon_Keeper(2769), OkirFun(2766), Kavish(2762), zaphus(2755), (1 more message) 2010-09-15T10:37:02 *** hornairs has joined #aichallenge 2010-09-15T10:37:27 aerique: posted the starter pack 2010-09-15T10:38:09 kool, i got TCP live game visualization now. 2010-09-15T10:38:12 some of it is stylistic/opinionated, no offense intended 2010-09-15T10:38:26 but I will miss fleets that last just one turn :( 2010-09-15T10:39:04 mega1: no problem, my style is usually "hey it compiles, ship it!" so the starter pack could use some work :) 2010-09-15T10:41:35 aerique: they should be compatible on the server because I kept MyBot.lisp around for the server to recognize that it should do 'make' which nets it a MyBot executable. 2010-09-15T10:44:16 mega1: server doesn't use make files 2010-09-15T10:44:45 the server cannot compile common lisp either 2010-09-15T10:45:06 dstufft: not trying to be arrogant here 2010-09-15T10:45:10 ... just being a natural. 2010-09-15T10:45:48 Any guidelines on how make it compilable? 2010-09-15T10:45:54 dstufft: it at least needs the info from the starter packs makefile, like in the tron challenge 2010-09-15T10:46:30 mega1: no worries, i don't know shit about lisp, I'm sure we'll add it once we get into the "add more languages" phase after the server stops getting stuck randomly 2010-09-15T10:46:52 aerique: ah, it was just copied from the Makefile to some build script? 2010-09-15T10:47:09 currently the system looks at the source file names 2010-09-15T10:47:15 to determine what command to run to build it 2010-09-15T10:47:27 ie if MyBot.cc is there, it builds it as c++ 2010-09-15T10:47:31 yes, that's why I left MyBot.lisp around 2010-09-15T10:47:48 although it's not even included in the final bot 2010-09-15T10:47:58 mega1: i think so, i'm not entirely sure how things ended up on the server last challenge. there's some reference to /usr/local/bin/make-clisp.sh in compile_all.py 2010-09-15T10:48:10 (filenames are not exact :-) ) 2010-09-15T10:48:36 oh well, I guess they shout when they get around to it. 2010-09-15T10:49:05 I modified compile_anything.py 2010-09-15T10:49:05 if it's part of compile_anything.py already likely the only thing we need to do is install any OS packages required 2010-09-15T10:49:09 so it is more flexible 2010-09-15T10:49:19 and add it to the languages db 2010-09-15T10:49:30 Bobng: does yours blow up randomly 2010-09-15T10:49:37 nope. 2010-09-15T10:49:39 yay 2010-09-15T10:49:52 compile_anything.py is fun, it likes to blow up on random shit 2010-09-15T10:49:59 http://code.google.com/p/ai-contest/issues/detail?id=130 2010-09-15T10:50:00 *** sinan has quit IRC (Quit: WeeChat 0.2.6.3) 2010-09-15T10:50:01 i'm slowly adding in checks for all the random shit though 2010-09-15T10:50:08 Have a look at my patch 2010-09-15T10:50:17 It cleans the code up a lot and makes it a lot more stable 2010-09-15T10:50:18 like including a directory named MyBot would cause the entire thing to just die 2010-09-15T10:50:24 and not able to compile anything 2010-09-15T10:50:29 from your submission on 2010-09-15T10:50:30 mega1: should've numbered your pack 0.4 :) 2010-09-15T10:50:51 mega1: i meant 0.3 2010-09-15T10:50:54 aerique: I left ample room for confusion. 2010-09-15T10:51:12 Well dstufft the code is a mess, i'm not surprised it blows up randomly 2010-09-15T10:51:20 i mean one function is 120+ lines long 2010-09-15T10:51:22 Bobng: yea it is 2010-09-15T10:51:26 i hate the backend code 2010-09-15T10:51:28 and the ww code 2010-09-15T10:51:33 well basically all the code 2010-09-15T10:51:33 I love the backend code 2010-09-15T10:51:36 heh 2010-09-15T10:51:37 aerique: by not "overtaking" the versions of the other cl packs 2010-09-15T10:52:01 but yeah, my patch should stop all the random blowing up issues if someone includes it. 2010-09-15T10:52:05 server down again? 2010-09-15T10:52:18 I want it to be included because i'm pretty proud of it, its got some nice abstractions 2010-09-15T10:52:29 Cyndre: compile daemon blew up 2010-09-15T10:52:36 i fixed it 50 minutes ago 2010-09-15T10:52:42 but it's still processing the back log 2010-09-15T10:52:54 Bobng: i'll take a look in a bit 2010-09-15T10:53:16 awsome 2010-09-15T10:54:16 needs some testing though 2010-09-15T10:54:30 but it shouldn't blow up. hundreds of if statements in a function is never a good thing -_- 2010-09-15T10:54:43 can it be worse than a current version? :) 2010-09-15T10:55:08 way to overengineer man. 2010-09-15T10:55:09 Naktibalda: sure --add another 100 if statments 2010-09-15T10:55:53 *** brox_ has joined #aichallenge 2010-09-15T10:56:09 hey ppl 2010-09-15T10:56:21 any admins over there? 2010-09-15T10:56:27 of course 2010-09-15T10:56:29 brox_: what do you need 2010-09-15T10:56:47 * edcba administers his own pc ! 2010-09-15T10:56:49 I forgot to state that we are a team, while registered 2010-09-15T10:57:16 what's a point of a team? 2010-09-15T10:57:35 and there are no fields to change now 2010-09-15T10:57:46 caytchen: if you call his patch overengineered, then i hate to see what you'd call the rewrite im doing of the entire system jsut for fun :p 2010-09-15T10:57:58 well, we have a fun as a team, whats wrong? 2010-09-15T10:58:31 *** kozlovsky has quit IRC (Ping timeout: 252 seconds) 2010-09-15T10:58:55 brox_: can you post on the forums, i don't know the rules re the teams stuff 2010-09-15T10:59:17 dstufft okie 2010-09-15T10:59:57 *** brox_ has left #aichallenge 2010-09-15T11:00:30 i really really hate svn 2010-09-15T11:00:34 just fyi 2010-09-15T11:02:22 *** caytchen has quit IRC (Ping timeout: 276 seconds) 2010-09-15T11:03:07 dstufft: me too 2010-09-15T11:03:24 everytime i find a problem, or something i want to work on 2010-09-15T11:03:30 with the ai server code 2010-09-15T11:03:34 i sit there and think 2010-09-15T11:03:37 It corrupted a repo for me once... shifted to git after that... never going back 2010-09-15T11:03:46 "well do i really feel like using svn right now" 2010-09-15T11:04:26 When is this ai challenge over? 2010-09-15T11:04:29 *** xtile has quit IRC (Quit: Page closed) 2010-09-15T11:05:30 3 months if i recall 2010-09-15T11:05:35 it should say on the homepage i think 2010-09-15T11:06:02 To true, write on the front page, nov 27 is the deadline. 2010-09-15T11:09:17 *** bss03 is now known as bss03-AFK 2010-09-15T11:09:30 danielvf is doing some great work 2010-09-15T11:09:35 in making the site not suck 2010-09-15T11:10:17 So It says we aren't allow to use threads, but what about using scala with an actor as our bot? 2010-09-15T11:10:36 *** aerique has quit IRC (Quit: ...) 2010-09-15T11:10:47 odinsbane: don't know shit about scala 2010-09-15T11:10:56 best bet is to post it on the forum 2010-09-15T11:10:57 Heh, this is interesting. The current incarnation of my bot does better against RageBot than against any of the other example bots. 2010-09-15T11:11:14 also yay 2010-09-15T11:11:18 we finally got through the backlog 2010-09-15T11:11:21 well 10 left 2010-09-15T11:11:31 so new submissions should be getting compiled pretty quikcly 2010-09-15T11:13:53 *** tobiassjosten has quit IRC (Quit: Leaving) 2010-09-15T11:13:59 odinsbane: no OS threads. green threads (that run in a single OS thread) are allowed 2010-09-15T11:14:29 we got the rankings stuff all nicely optimized now. thanks to danielvf! 2010-09-15T11:17:11 *** TapewormGalactos has joined #aichallenge 2010-09-15T11:18:02 Hi, I keep losing games, but the replay shows the opposite and I don't know why... maybe somebody can explain: http://ai-contest.com/visualizer.php?game_id=4479822 2010-09-15T11:19:15 TapewormGalactos: my best guess is that your bot is doing something illegal as soon as it sees that the game is in the bag 2010-09-15T11:19:22 TapewormGalactos: do you have anything that casues a crash when you "win" 2010-09-15T11:19:34 I don't think so 2010-09-15T11:19:38 what could that be? 2010-09-15T11:19:44 ie enemy_planets[0] (there are no enemy_plants) 2010-09-15T11:19:47 or things of that nature 2010-09-15T11:20:00 for example, it's easy to just stop playing when it's clear that you win sometimes 2010-09-15T11:20:08 one version of my bot last year did that 2010-09-15T11:20:08 Enemy pants 2010-09-15T11:20:14 hm I'm gonna look into that 2010-09-15T11:20:28 *last contest 2010-09-15T11:20:48 ENEMY PLANTS 2010-09-15T11:20:57 planet wars has turned into Weed Wars 2010-09-15T11:21:07 i'll be modifying the server shortly 2010-09-15T11:21:53 add more species 2010-09-15T11:27:07 *** boegel has quit IRC (Quit: Leaving) 2010-09-15T11:30:46 *** coventry has quit IRC (Ping timeout: 240 seconds) 2010-09-15T11:31:15 *** l4u has joined #aichallenge 2010-09-15T11:32:22 *** hornairs has quit IRC (Quit: hornairs) 2010-09-15T11:34:32 *** Kastra has joined #aichallenge 2010-09-15T11:35:39 use thepriceofweed.com for location based princing 2010-09-15T11:36:23 http://www.priceofweed.com/ sorry wrong link 2010-09-15T11:36:32 oh nice 2010-09-15T11:36:55 nice 2010-09-15T11:38:03 no weed in UK? 2010-09-15T11:38:20 *** delt0r_ has joined #aichallenge 2010-09-15T11:39:07 so when using the java visualization, when you time out, does that mean you used more than 1 sec to compute? 2010-09-15T11:40:11 *** delt0r___ has quit IRC (Ping timeout: 265 seconds) 2010-09-15T11:42:57 *** askhader has quit IRC (Ping timeout: 240 seconds) 2010-09-15T11:43:06 *** Racko has quit IRC (Read error: Connection reset by peer) 2010-09-15T11:43:39 *** askhader has joined #aichallenge 2010-09-15T11:45:09 rank me.... http://ai-contest.com/profile.php?user_id=3911 2010-09-15T11:45:10 *** amstan has quit IRC (Remote host closed the connection) 2010-09-15T11:54:46 *** bduc has quit IRC (Ping timeout: 255 seconds) 2010-09-15T12:00:41 *** bduc has joined #aichallenge 2010-09-15T12:00:56 *** damag has joined #aichallenge 2010-09-15T12:01:07 TapewormGalactos, or there was an unhandled exception of somesort 2010-09-15T12:01:29 for example, if you have a planet with 100 ships and you order a move of 120 ships from that planet, you timeout and die 2010-09-15T12:01:42 yay I got a game 2010-09-15T12:02:58 rogue780: I see, that might be it 2010-09-15T12:04:06 TapewormGalactos: have you the same issue against example bots? 2010-09-15T12:04:13 yes 2010-09-15T12:04:21 sometimes 2010-09-15T12:04:24 log all your orders and check the last one 2010-09-15T12:05:12 *** Vi0 has joined #aichallenge 2010-09-15T12:08:51 I'll try that 2010-09-15T12:11:47 *** Kastra2 has joined #aichallenge 2010-09-15T12:12:22 *** Kastra has quit IRC (Ping timeout: 240 seconds) 2010-09-15T12:17:44 *** HakanD has joined #aichallenge 2010-09-15T12:18:21 *** justin_pdx has joined #aichallenge 2010-09-15T12:20:21 *** Kastra2 is now known as Kastra 2010-09-15T12:22:10 *** jukkerknott has joined #aichallenge 2010-09-15T12:23:53 *** jukkerknott has quit IRC (Client Quit) 2010-09-15T12:27:47 *** phreeza has joined #aichallenge 2010-09-15T12:32:39 *** McLeopold has joined #aichallenge 2010-09-15T12:32:51 http://www.benzedrine.cx/planetwars/canvas?game_id=1284566256|deepblue|McLeopold.5 2010-09-15T12:33:01 I'm amazed at this game. 2010-09-15T12:34:38 I think you would have lost that on the official server 2010-09-15T12:34:48 how's that? 2010-09-15T12:34:51 Since it's limited to 200 turns 2010-09-15T12:35:04 And it seems like you had less ships on turn 200 2010-09-15T12:35:06 oh, is that official yet? 2010-09-15T12:35:35 I'm not sure, but it's been that way for a while 2010-09-15T12:35:37 It's less impressive if deepblue just stops trying at 200. 2010-09-15T12:36:08 Ya I'm wondering if his logic changes based on turn number or something 2010-09-15T12:36:16 But either way, good job =) 2010-09-15T12:36:20 nice match 2010-09-15T12:37:56 *** amstan has joined #aichallenge 2010-09-15T12:37:56 *** ChanServ sets mode: +o amstan 2010-09-15T12:39:04 It's hard to tell, but round 200 is where it started to turn. 2010-09-15T12:39:19 Right, that's why it made me think that =) 2010-09-15T12:39:59 *** dhartmei has quit IRC (Quit: brb) 2010-09-15T12:40:32 *** dhartmei has joined #aichallenge 2010-09-15T12:42:31 *** num1_ has joined #aichallenge 2010-09-15T12:43:46 *** maiden has joined #aichallenge 2010-09-15T12:43:52 *** sugarshark has joined #aichallenge 2010-09-15T12:44:22 gah i thought the server had an old version of ghc 2010-09-15T12:44:31 devio.us (openbsd shell) has ghc 6.6.1 2010-09-15T12:44:45 *** amstan has quit IRC (Ping timeout: 265 seconds) 2010-09-15T12:45:04 i bet it can't even load my haskell starter 2010-09-15T12:45:26 lol "PlanetWars.lhs:21:14: cannot parse LANGUAGE pragma" 2010-09-15T12:45:45 i kind of expected that part to work but just not support the extension 2010-09-15T12:46:18 can some one make a step by step guide to how one is supposed to host a server. we want to run an internal turnament and do not wish to wait for houers between maches.. atleast an readme with some hints... 2010-09-15T12:47:08 maiden: i think it's Janzert1 who has created a guide to setting up a clone of the official server somewhere. it's somebody whose name starts with a J anyway 2010-09-15T12:47:12 smr: bruteforcing parameters on benzedrine ? 2010-09-15T12:47:36 smr: significant answers require dozens of games, maybe hundreds :) 2010-09-15T12:47:51 oh nice! so there is a guide :) 2010-09-15T12:48:23 bartwe_: he's already played almost a hundred for each parameter setting 2010-09-15T12:48:41 nice 2010-09-15T12:49:09 the spread is miniscule 2010-09-15T12:49:29 well, smr2-25-7 is clearly bad 2010-09-15T12:49:34 hehe 2010-09-15T12:49:48 I wonder if I change my handle to Sarah or Jenny if I'll get faster and less sarcastic responses in the different language specific channels 2010-09-15T12:49:54 what;s about thwse smr2 bots? 2010-09-15T12:49:57 maiden: don't forget the existence of http://www.benzedrine.cx/planetwars/ if that's appropriate 2010-09-15T12:50:11 rogue780: did you ever get your htdocs working? 2010-09-15T12:50:15 rogue780: you will get sexists responses 2010-09-15T12:50:20 sexist 2010-09-15T12:50:48 maiden, and if you need a pre-compiled tcp client for windows for the server at benzedrine, I put one together at www.errgaming.com/tcp.zip 2010-09-15T12:50:48 Nak: there are alot of them, and the names suggest that the onwer is trying out two parameter spaces 2010-09-15T12:50:56 McLeopold, I haven't tried again yet 2010-09-15T12:50:59 you? 2010-09-15T12:51:17 what kind of parameters? 2010-09-15T12:51:26 it doesn;t say 2010-09-15T12:51:28 No, I started programming my bot after benz was back up 2010-09-15T12:51:57 I got the server working, but I'm not sure about results files. 2010-09-15T12:51:59 McLeo: good bot btw 2010-09-15T12:52:19 McLeopold, same here. people can play on it and see if they win or lose, but that's about it 2010-09-15T12:52:20 Thanks, I credit it to all my loses against you. 2010-09-15T12:52:32 you seem to be winning recently 2010-09-15T12:52:46 A fake win against deepblue. 2010-09-15T12:52:58 does greghaynes's qt client do a visualiztion of the game? If so, combining it with our servers might actually make them useful 2010-09-15T12:53:04 i don't know that deepblue really gave up there. 2010-09-15T12:53:17 McLeopold: yeah it looks like deepblue started dropping in numbers within one turn of 200 2010-09-15T12:53:18 he had a numbers advantage but not a growth advantage just prior to 200 2010-09-15T12:53:20 ok. but i am trying to start an replica of the current server; with web interface and all. it will have more then 20 users. 2010-09-15T12:53:33 looks like a giveup to me, personally 2010-09-15T12:53:46 there was a critical abandonment of a 5 planet in his heartland 2010-09-15T12:54:04 keeping in mind that deepblue has a search algorithm which probably uses the 200 turn limit to keep things a bit more realistic 2010-09-15T12:54:42 maiden: yeah Janzert1's thing would help if you can get a hold of him 2010-09-15T12:54:44 Yea, I have no turn limit logic. I was waiting for official word. 2010-09-15T12:55:09 i will have turn limit dependent code. i'll just keep changing it until the number is fixed somewhere 2010-09-15T12:55:39 and probably will have a different version of my bot for benzedrine.cx than from the official server 2010-09-15T12:56:29 That way when you do submit, you can say "I'm not left handed!" 2010-09-15T12:57:14 sorry for such a stupid question, but how would a turn limit impact your bot's decision making? 2010-09-15T12:57:20 *** amstan has joined #aichallenge 2010-09-15T12:57:20 *** ChanServ sets mode: +o amstan 2010-09-15T12:57:23 hello 2010-09-15T12:57:28 @seen j3camero 2010-09-15T12:57:28 amstan: j3camero was last seen in #aichallenge 14 hours, 6 minutes, and 28 seconds ago: 967,345,723 2010-09-15T12:57:32 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716]) 2010-09-15T12:57:47 rogue780: it does if you're optimizing an objective function which is "number of my ships at turn 200 divided by number of my opponent's ships at turn 200" 2010-09-15T12:57:49 @seen Accoun 2010-09-15T12:57:49 amstan: Accoun was last seen in #aichallenge 3 hours, 37 minutes, and 9 seconds ago: whwre is source of PlayGame.jar ? 2010-09-15T12:58:02 a1k0n, ah, that makes sense 2010-09-15T12:58:08 jmcarthur: what's known about deepblue's search algorithm? 2010-09-15T12:58:22 it occurs to me that you might as well just optimize to a horizon equal to the map's diameter (in turns) though 2010-09-15T12:58:24 mcleo: http://www.benzedrine.cx/planetwars/canvas?game_id=1284567569|bartwe.7.8.1|McLeopold.5 any hints on how to beat you here ? 2010-09-15T12:58:27 its very strong 2010-09-15T12:58:29 mega1: just that it's a search algorithm, afaik 2010-09-15T12:58:34 bah 2010-09-15T12:59:10 a1k0n: but what if you bot plans to obtain enough ships somewhere between 200 turns and your horizon metric? 2010-09-15T12:59:15 *your 2010-09-15T12:59:57 actually i don't think i understood your proposal 2010-09-15T13:00:10 Sometimes I think it's luck. 2010-09-15T13:00:23 McLeo: it looks good 2010-09-15T13:00:44 it smells like neutral stealing 2010-09-15T13:00:52 not sure why i don't return the favor there 2010-09-15T13:00:59 haskell people: this is completely untested, but... http://jake.devio.us/starter/PlanetWars.html 2010-09-15T13:01:01 Janzert1 are you here? 2010-09-15T13:01:01 too strong protection logix i guess 2010-09-15T13:01:42 i'm hoping it doesn't have too many bugs 2010-09-15T13:02:00 the parsing and IO is tested (a little). the game engine logic is not at all tested 2010-09-15T13:02:23 dang, i don't export enough stuff 2010-09-15T13:02:30 okay i have a few more changes to make :) 2010-09-15T13:02:38 jmcarthur: i haven't thought through it yet. heh. i'm going to just shut up and get a bot together. 2010-09-15T13:03:01 i have something in mind now, but i worry it'll be way too conservative. we'll see. 2010-09-15T13:03:19 yeah, i still don't really have a working bot. just drafts of ideas, basically 2010-09-15T13:03:23 bartwe_: I had a 60 ship advantage at turn 27. I think that made the difference. 2010-09-15T13:03:32 this extending time period has slowed me down 2010-09-15T13:03:37 *extended 2010-09-15T13:03:42 not that that's a bad thing 2010-09-15T13:03:48 it's given me some time to breathe 2010-09-15T13:04:20 McLeo: yep, still counting where you got that advantage 2010-09-15T13:04:42 what's the new time limit? it's always been 200 right? 2010-09-15T13:05:11 it's 300 on benzedrine.cx 2010-09-15T13:05:13 num1_: it was 100 before, now it's 200, benzedrine uses 300, none of it is concrete 2010-09-15T13:05:19 Only way to create a difference in fleet is to attack neutrals and to have a different growth rate 2010-09-15T13:06:11 bartwe_: you got to the 14 planet first and got 3 ships up on me at 12... 2010-09-15T13:06:34 yep 2010-09-15T13:06:39 bartwe_: the small 14 planet 2010-09-15T13:06:41 *** hornairs has joined #aichallenge 2010-09-15T13:06:42 Ok, I have a first bot ready. It's written in clojure, meaning it runs as a jar file. It's more than 2 megabyte (close to 4) due to having clojure embedded. Anybody know if I can still play? 2010-09-15T13:06:59 bartwe_: I spent time annoying your 5 growth planet. 2010-09-15T13:07:55 step 16 i guess was critical 2010-09-15T13:08:21 going from -3 to +7 2010-09-15T13:08:27 it's so hard to analyze these compared to tron 2010-09-15T13:09:16 intuition seems to work for the simpler tactical bits 2010-09-15T13:09:26 bartwe_: If you defend better, then you don't have ships for attack. It's a tradeoff. 2010-09-15T13:09:43 defense is great when you have the upper hand already 2010-09-15T13:09:44 McL: it should have been a standoff 2010-09-15T13:09:52 especially if you have better growth 2010-09-15T13:09:58 defense is not so great if you are behind anyway 2010-09-15T13:10:08 defense is useless if your behind 2010-09-15T13:10:27 i wouldn't say useless. just don't focus on it 2010-09-15T13:10:28 but whats defense but the capturing of a future enemy planet ? 2010-09-15T13:10:42 i never programmed any defense 2010-09-15T13:10:46 if it means the difference between keeping and losing a 5-growth, you should hold on to it 2010-09-15T13:11:10 the downside to defense is that it's hard to counter-attack in this game 2010-09-15T13:11:23 war is zerosum in this thing 2010-09-15T13:11:33 there is so much time to prepare for an incoming attack 2010-09-15T13:11:36 its more about exposure 2010-09-15T13:11:55 growthrate is everything 2010-09-15T13:11:59 absolutely 2010-09-15T13:12:05 http://www.benzedrine.cx/planetwars/canvas?game_id=1284569574|McLeopold.5|bartwe.7.8.1 2010-09-15T13:12:14 bartwe_: win # 2 2010-09-15T13:13:08 hmz, no loading on this mobile internet 2010-09-15T13:13:14 not 2010-09-15T13:15:20 what kind of ai are you using ? 2010-09-15T13:15:53 Microsoft BoB 2010-09-15T13:15:57 smart 2010-09-15T13:16:03 i bet nobody else does that 2010-09-15T13:16:04 *** amstan_ has joined #aichallenge 2010-09-15T13:16:04 *** ChanServ sets mode: +o amstan_ 2010-09-15T13:16:17 *** RainCT has joined #aichallenge 2010-09-15T13:16:20 har 2010-09-15T13:17:24 I really limitted my calculations at first because of a npe, running the game locally showed it as a timeout. 2010-09-15T13:17:49 yay 2010-09-15T13:17:57 powerloss in MC 2010-09-15T13:18:00 *** pgpaskar_ has quit IRC (Remote host closed the connection) 2010-09-15T13:18:13 *** oh_cripes has quit IRC (Ping timeout: 252 seconds) 2010-09-15T13:18:15 MC ? 2010-09-15T13:18:19 yay i have bugs in my starter 2010-09-15T13:18:21 *** cbeck has left #aichallenge 2010-09-15T13:18:30 It's a secret... 2010-09-15T13:18:39 funny how i managed to do that even right next to the spec 2010-09-15T13:18:49 *** Naktibalda has joined #aichallenge 2010-09-15T13:19:06 people aren't as forthcoming in this contest as the last one, it seems 2010-09-15T13:19:08 i guess i didn't take the literate programming approach very strictly. i didn't actually address my code in the text. the text is basically just a split up version of the spec 2010-09-15T13:19:10 *** amstan has quit IRC (Ping timeout: 240 seconds) 2010-09-15T13:19:11 *** askhader has quit IRC (Remote host closed the connection) 2010-09-15T13:19:11 *** j3camero has quit IRC (Remote host closed the connection) 2010-09-15T13:19:22 a1k0n: yeah they all have different strategies this time, i assume! 2010-09-15T13:19:34 i too am hesitant to reveal my idea 2010-09-15T13:19:45 a1kon: you upped the competitiveness with the promise of more intresting jobs 2010-09-15T13:19:49 i don't have enough ideas to reveal 2010-09-15T13:19:50 heh 2010-09-15T13:20:00 haha 2010-09-15T13:20:06 more interesting jobs? 2010-09-15T13:20:10 i mean, i had a really bad idea earlier 2010-09-15T13:20:17 lol 2010-09-15T13:20:21 there's like 20 ppl left in csc 2010-09-15T13:20:23 Naktibalda: a1k0n won the last contest and got tons of recruiters chasing him down 2010-09-15T13:20:32 It's kinda fun this way. Afterward a suppose we can all collaborate for the best bot ever. 2010-09-15T13:20:41 which ended up being the equivalent of assuming there'd be one fleet per turn per planet 2010-09-15T13:20:50 idea #1, don't let your bot crash, that should get you pretty good rankings on ai-contest.com. 2010-09-15T13:20:57 i never did find out what hebbie was doing with this tron bot which eventually destroyed mine after the contest on benzedrine.cx 2010-09-15T13:21:10 probably overfit to your bot :P 2010-09-15T13:21:27 exploited just one particular flaw or something 2010-09-15T13:21:30 Janzert1's bot was also way, way better at the max-path-length problem 2010-09-15T13:21:57 my bot had to assume everyone else was as bad as it was at that 2010-09-15T13:21:57 my bot was erratic when it came to filling spaces 2010-09-15T13:22:11 except when it was about halfway done 2010-09-15T13:22:20 then it would be like "oh crap i've been screwing up" and try to recover 2010-09-15T13:22:22 the problem with alpha-beta is you can't get away with underestimating/overestimating asymmetrically 2010-09-15T13:22:29 heh 2010-09-15T13:22:30 mine did worse, disqualified for failing to enter the contest 2010-09-15T13:22:46 aw 2010-09-15T13:22:50 ha. 2010-09-15T13:22:52 i failed to enter the contest, too! 2010-09-15T13:22:58 because i didn't hear about it :P 2010-09-15T13:23:10 failure buddies 2010-09-15T13:23:21 i thought mine was interesting for having a completely different approach from most others, but i never worked out all the kinks 2010-09-15T13:23:26 bartwe_: you didn't know about the tron contest? i guess we had previously corresponded on irc re: mario ai and icfp? 2010-09-15T13:23:33 mine would win where others wouldn't, and lose where others would have probably won 2010-09-15T13:23:34 yep 2010-09-15T13:23:47 yeah the tron thing intersected with moving houses 2010-09-15T13:23:55 *** amstan_ is now known as amstan 2010-09-15T13:24:04 ah. yeah i just moved across the country, and *then* this contest started 2010-09-15T13:24:29 great way to start a new job, huh? 2010-09-15T13:24:44 yeah. luckily my boss is thinking of entering, himself 2010-09-15T13:24:54 haha 2010-09-15T13:24:59 i'm a father for 6 weeks now, that also is not a help in the time department 2010-09-15T13:25:31 ah yes, that has been an issue for me. last night my 2yo just wanted to build jigsaw puzzles over and over and over 2010-09-15T13:25:55 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-09-15T13:26:11 *** iFire has joined #aichallenge 2010-09-15T13:26:19 *** Cyndre has quit IRC (Read error: Connection reset by peer) 2010-09-15T13:26:35 *** Cyndre has joined #aichallenge 2010-09-15T13:26:50 but at least i don't have to wake up every three hours now. man the first couple months are hard 2010-09-15T13:26:52 Hey, I'm a new dad too. My son is 6m. 2010-09-15T13:27:11 hehe 2010-09-15T13:27:24 young fathers GO! 2010-09-15T13:27:27 I installed babysmash for him yesterday. He loves it. 2010-09-15T13:27:41 i just realized something about this game 2010-09-15T13:27:48 time == space 2010-09-15T13:28:01 yes, i measure the diameter of the universe in turns 2010-09-15T13:28:07 fleets travel at the speed of light 2010-09-15T13:28:19 or the speed of fleets anyway 2010-09-15T13:28:37 space is also slightly convex 2010-09-15T13:28:40 well, and the planets are $dist light-turns apart 2010-09-15T13:28:51 also has implications for determining threat to a planet 2010-09-15T13:28:54 light-turns 2010-09-15T13:28:56 heh 2010-09-15T13:28:59 *** Palmik has quit IRC (Remote host closed the connection) 2010-09-15T13:29:07 you just realized this now, and your bot is that good? 2010-09-15T13:29:11 *** Palmik has joined #aichallenge 2010-09-15T13:29:31 a1kon: i look into the future 2010-09-15T13:29:39 a1kon: i should also be looking into the past 2010-09-15T13:29:55 *** yuwei has quit IRC (Ping timeout: 265 seconds) 2010-09-15T13:30:17 read steven hawkings take on cones of influence 2010-09-15T13:30:29 hmz, i think i have that somewhere 2010-09-15T13:30:29 *** yuwei has joined #aichallenge 2010-09-15T13:31:13 ha lets turn this bot into a searcher :) 2010-09-15T13:32:01 *** hornairs has quit IRC (Quit: hornairs) 2010-09-15T13:33:20 *** amstan has quit IRC (Ping timeout: 255 seconds) 2010-09-15T13:33:45 *** rwa has quit IRC (Quit: rwa) 2010-09-15T13:34:10 *** drewrobb_ has joined #aichallenge 2010-09-15T13:36:06 a1kon: 'this good' is going to count for very little in a few weeks 2010-09-15T13:36:27 *** caytchen has joined #aichallenge 2010-09-15T13:36:29 *** Yoshi-TS4 has quit IRC (Ping timeout: 265 seconds) 2010-09-15T13:38:14 *** drewrobb has quit IRC (Ping timeout: 276 seconds) 2010-09-15T13:40:29 Is there any variation in the startup of a map, ie two bots that start on the same map, should return the same outcome unless somebody has written their bot to be a bit random? 2010-09-15T13:41:21 odinsbane: the engine and maps are completely deterministic 2010-09-15T13:41:22 ive added random precsiely for that reason 2010-09-15T13:41:54 *** troutwine has joined #aichallenge 2010-09-15T13:41:59 odinsbane: only the time limits and turn counts are unspecified, to my knowledge, but given that those are not factors, it should be the same from run to run 2010-09-15T13:42:25 and i would like for the turn counts to be specified 2010-09-15T13:42:28 *** justin_pdx has joined #aichallenge 2010-09-15T13:42:34 Just to be on the safe side, our submitted package can contain any code so long as it compiles into MyBot, right? 2010-09-15T13:42:50 So does anybodies bot here beat all of the starter bots on all of the maps? 2010-09-15T13:42:57 troutwine: and so long as the compilation process the server performs will build it properly 2010-09-15T13:43:05 I would hope everyone does... 2010-09-15T13:43:06 odinsbane: a lot of peoples' do 2010-09-15T13:43:06 odinsbane, Yea. 2010-09-15T13:43:09 *** maiden_ has joined #aichallenge 2010-09-15T13:43:11 not everyone 2010-09-15T13:43:20 all maps all bots... thats just for debug 2010-09-15T13:43:35 jmcarthur: Great, thanks. 2010-09-15T13:44:33 what kind of algorithm does bartwe use for that infernal chain of death? 2010-09-15T13:44:34 actaully.... i haven't tested 2010-09-15T13:44:43 rogue: a simple one 2010-09-15T13:45:21 *** maiden has quit IRC (Ping timeout: 252 seconds) 2010-09-15T13:45:37 rogue780: i don't know what bartwe_ uses, but it looks like a minimal spanning tree to me 2010-09-15T13:45:42 basicly after alot of other steps the 'free' population of a planet is decided, if it is above a threshold compared to its 'parent' planet, it sends it to the parent 2010-09-15T13:46:49 maybe is should make it a loop 2010-09-15T13:46:49 vortex 2010-09-15T13:46:56 ha 2010-09-15T13:46:59 that would be scary 2010-09-15T13:47:09 dear god no 2010-09-15T13:47:24 it already resembles the death start focusing its beams onto a center point before firing big 2010-09-15T13:47:28 *death star 2010-09-15T13:47:36 hopefully I just fixed my bot's "oh god, i'm being attacked so I shouldn't send any fleets" bug 2010-09-15T13:47:38 ha ha 2010-09-15T13:48:18 rogue: actually, losing fleets is a positive move if your in a numerical advantage 2010-09-15T13:49:14 if by "losing fleets" you mean "investing fleets" 2010-09-15T13:49:32 *** Theobon has joined #aichallenge 2010-09-15T13:49:45 *** maiden_ has quit IRC (Ping timeout: 252 seconds) 2010-09-15T13:49:53 indeed 2010-09-15T13:49:58 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-15T13:50:02 I just can't seem to get that advantage lately. 2010-09-15T13:50:11 meh, it makes the advantage relatively larger compared to total population 2010-09-15T13:50:19 *** amstan has joined #aichallenge 2010-09-15T13:50:19 *** ChanServ sets mode: +o amstan 2010-09-15T13:50:24 still, location, location, location 2010-09-15T13:50:31 hence the time == space thing 2010-09-15T13:50:37 like in the last match we just played. you destroyed my starting move 2010-09-15T13:51:00 meh, i lose matches in move 10 too 2010-09-15T13:51:06 or move 2 2010-09-15T13:51:14 or 1 :) 2010-09-15T13:51:33 hard to know how much to invest in neutrals 2010-09-15T13:52:10 Especially if you wait, then your opponent might invest in them for you. 2010-09-15T13:52:44 or your one step behind and start nicely behind in population, from which it is hard to get back 2010-09-15T13:53:34 *** HakanD has quit IRC (Ping timeout: 240 seconds) 2010-09-15T13:55:04 how do you calculate the time for a neutral to became investment neutral ? 2010-09-15T13:55:17 har 2010-09-15T13:56:00 *** McLeopold has quit IRC (Quit: Leaving) 2010-09-15T13:57:05 I just sorta smear distance/numships and growth into a number. 2010-09-15T13:57:14 t = neutralforce div growth is what it takes to get neuytral 2010-09-15T13:57:48 but that means both a fleet 'invested for distance +- t and that the enmey has a lower cost to take it 2010-09-15T13:57:49 so maybe t = distance + neutral * 2 / growth ? 2010-09-15T13:57:51 does enybody use unit tests for bot programming? 2010-09-15T13:57:56 anybody 2010-09-15T13:58:20 *** num1_ has quit IRC (Ping timeout: 276 seconds) 2010-09-15T13:58:37 a planet would be a great take oif t is lower then the soonest enemy attack 2010-09-15T13:58:43 Nak: not me 2010-09-15T13:59:43 i would test properties of my code if i actually took the time to devise a simple model for it 2010-09-15T14:00:15 (which i may actually do) 2010-09-15T14:00:33 I was trying to thing of it as a return. How many turns does it take before their is a yield. 2010-09-15T14:00:52 exactly one, its growth. 2010-09-15T14:00:56 bartwe_: where do you get the factor of 2 from? Does that have to do if it is in enemy hands? 2010-09-15T14:01:09 my bot will take a lot of upfront investment in terms of data structures, but my idea has the benefit that i don't have to think about such things as the cost of taking a neutral 2010-09-15T14:01:16 or at least i don't have to think so deeply about it 2010-09-15T14:01:22 odinsbane: that a fudge factor, its higher then 1 atleast, not sure yet how to determine it 2010-09-15T14:01:59 neutral have a lower cost if you know that the enemy can't simply take it after you invest 2010-09-15T14:02:13 that is, neutral costs less if you are also attacking the enemy as you expand 2010-09-15T14:02:21 odins: its a factor to ensure that t is high enough to make an enemy snipe value neutral or positive 2010-09-15T14:02:25 because that makes it harder for them to take your neutral 2010-09-15T14:03:25 you could deduct from tby sending a bigger force if you have superior numbers 2010-09-15T14:03:40 my bot is doing alot better now :) 2010-09-15T14:03:41 assuming that supresses an initial snipe 2010-09-15T14:04:00 closed-form solutions are nice, but not likely in this case 2010-09-15T14:04:04 bartwe_: so do you handle that more carefully now? 2010-09-15T14:04:05 http://www.benzedrine.cx/planetwars/getplayer?player=Cyndre 2010-09-15T14:04:07 How about calculations, do you limit yourself based on time? It seems like 1 second is a long time. 2010-09-15T14:04:35 the time limit is going to be my biggest hurdle 2010-09-15T14:04:51 caythen: haven't implemented this yet 2010-09-15T14:05:11 just thinking about heuristics for a decision 2010-09-15T14:05:31 i do need a better way to determine whcih neutral to attack 2010-09-15T14:05:39 if any 2010-09-15T14:05:55 especially knowning which neutral not to attack would be good :) 2010-09-15T14:06:09 *** smhanov_ has joined #aichallenge 2010-09-15T14:06:44 *** deepblue has joined #aichallenge 2010-09-15T14:06:55 *** num1_ has joined #aichallenge 2010-09-15T14:07:25 The final time limit could make a difference for that. 2010-09-15T14:08:17 do you have 3 + 1*199 seconds in the sandbox ? 2010-09-15T14:08:20 Cyndre: is it just me or do you totally overkill on incoming enemy fleets? 2010-09-15T14:08:24 or am i free to start multiple threads ? 2010-09-15T14:10:05 *** smhanov__ has quit IRC (Ping timeout: 276 seconds) 2010-09-15T14:12:07 *** eburnette has joined #aichallenge 2010-09-15T14:12:18 man those capthas are getting hard :) 2010-09-15T14:12:31 capthas ? 2010-09-15T14:12:33 *** phreeza has quit IRC (Quit: Computer has gone to sleep) 2010-09-15T14:12:37 to get into the chat room 2010-09-15T14:12:51 cool, ascii captchas? 2010-09-15T14:12:57 man irc has advanced ;p 2010-09-15T14:13:09 web interface, http://webchat.freenode.net/?channels=aichallenge&uio=d4 2010-09-15T14:13:18 well, shit. the more I try and improve my bot, the worse is seems to get 2010-09-15T14:13:20 ow 2010-09-15T14:13:41 rogue: inverted input to a cost function ? 2010-09-15T14:13:51 bug in prediction ? 2010-09-15T14:13:53 hey guys, is there a way to tell what map was used for a game, for example http://www.benzedrine.cx/planetwars/canvas?game_id=1284572451|deepblue|McLeopold.5 2010-09-15T14:14:37 map_id=maps/map80.txt 2010-09-15T14:14:46 check the source of the page itself for the playback data 2010-09-15T14:15:04 great. thanks Zannick for a very nice service 2010-09-15T14:15:07 buggy version of my bot :) 2010-09-15T14:15:21 even if you don't have the map itself, the front of the playback string has it 2010-09-15T14:15:24 bartwe i'm not sure where the problem lies. 2010-09-15T14:15:26 *** boegel has joined #aichallenge 2010-09-15T14:15:31 eburnette: um, you're welcome, but it's not my server 2010-09-15T14:15:50 *** bix0r has joined #aichallenge 2010-09-15T14:16:09 caytchen: no, I totally overkill - thats the next part to rewrite (once the bot is making more ships it uses a different piece of code - only updated the part that helps it get in front) 2010-09-15T14:16:19 Ah, thanks to dhartmei then :) 2010-09-15T14:17:40 *** hellman_ has joined #aichallenge 2010-09-15T14:19:08 *** hellman has quit IRC (Ping timeout: 276 seconds) 2010-09-15T14:19:29 *** amstan has quit IRC (Ping timeout: 264 seconds) 2010-09-15T14:19:44 What's the time limit for compiling a bot? 2010-09-15T14:23:56 gah. for some reason I keep getting zeros on my scores 2010-09-15T14:25:01 hwo fast do fleets travel? 1 distance unit per turn? 2010-09-15T14:25:15 rogue780: check the games being played 2010-09-15T14:25:27 distance is measured in turns it takes 2010-09-15T14:25:42 Zannick: great, thx, makes perfect sense :) 2010-09-15T14:25:53 fleets travel at a rate of 1 turn per turn 2010-09-15T14:25:54 :) 2010-09-15T14:25:58 boegel, I was refering to my bot's decision makning scoring process. 2010-09-15T14:26:35 rogue780: ah :) 2010-09-15T14:26:42 rogue780: rounding error then? :P 2010-09-15T14:27:35 integer division? 2010-09-15T14:28:51 I'm getting NaN for this: (double)pw.Distance(e.PlanetID(), p.PlanetID())/ Math.pow((double)pw.Distance(e.PlanetID(), p.PlanetID()), 2); where e is an enemy planet and p is a neutral planet 2010-09-15T14:29:38 rogue780: did you check if they're the same planet? 2010-09-15T14:29:44 How come some planets have 0 growth rate? Is that normal? 2010-09-15T14:30:06 eburnette: yep, intended 2010-09-15T14:30:23 odinsbane, yes. they're not the same 2010-09-15T14:30:37 interesting, can you think of a reason why i'd want to take over a 0 growth planet? 2010-09-15T14:30:52 The zero growth planets were confusing to me at first. On the viewer they don't show up, they're just a black hole with a population. 2010-09-15T14:30:59 *** amstan has joined #aichallenge 2010-09-15T14:30:59 *** ChanServ sets mode: +o amstan 2010-09-15T14:31:02 eburnette: to obtain more planet than your enemy? 2010-09-15T14:31:51 you would only ever want to attack one to kill an enemy fleet stationed on it 2010-09-15T14:31:52 On map1 in the demo packages there is a zero growth planet smack in the middle of the map with a ring around it. It might be a strategic place to pile up ships since they could go anywhere on the map faster than any other place on the map. 2010-09-15T14:32:03 ^^ 2010-09-15T14:33:15 bartwe_: the spec is incorrect (regarding the 3 second first turn). i have submitted a patch for it 2010-09-15T14:33:20 bartwe_: each turn is only 1 second, in reality 2010-09-15T14:33:31 I was thinking of unowned zero growth planets, justin_pdx that makes sense 2010-09-15T14:33:47 is there a link the greghaynes visualizer/tcp server client 2010-09-15T14:33:55 anyway, i gotta take some trash to the dump and then mow the lawn. i'll be back in a few hours 2010-09-15T14:33:57 Cyndre: search for greghaynes on github 2010-09-15T14:34:01 But that would take geographic strategic reasoning to evaluate the worth of a zero growth planet and I'm no where near that. I'm still trying to figure out how to maximize growth without thought of winning. 2010-09-15T14:34:10 the greghaynes tcp client doesn't work on windows AFAIK 2010-09-15T14:34:28 is there one that does? 2010-09-15T14:34:36 is you port it it does 2010-09-15T14:34:39 *if 2010-09-15T14:34:56 Has anybody set up a client so you (a human) can try your hand against a bot? 2010-09-15T14:35:02 it's a qt app. qt works on windows. you probably just have to change something fairly minor 2010-09-15T14:35:07 eburnette, Yea 2010-09-15T14:35:11 I was thinking of making a flash or an html5/nodejs client 2010-09-15T14:35:14 Its totally useless 2010-09-15T14:35:30 Cyndre: how does it run your bot? 2010-09-15T14:35:35 unless you run with fixed turns --and hit the "next" turn thing 2010-09-15T14:35:51 jmcarthur: just a human interface to the tcp server 2010-09-15T14:36:08 delt0r, real time 1sec per turn is fine 2010-09-15T14:36:13 delt0r: or monitor time and end turn every second and the human just dispatch's 2010-09-15T14:36:25 trust me--even simple bots completly own you 2010-09-15T14:36:31 I can see that 2010-09-15T14:36:41 *** fx4m has quit IRC (Ping timeout: 252 seconds) 2010-09-15T14:36:48 maybe, would be fun to try 2010-09-15T14:37:09 you simply can't click that often without breaking your mouse and also being a starcraft god 2010-09-15T14:37:35 well I intend to release the code...but probalby be the weekend 2010-09-15T14:37:46 You'd need a Starcraft apm of 1000+. 2010-09-15T14:37:54 its java... but any bot can be run with it 2010-09-15T14:38:09 and Its lets you watch tcp games too 2010-09-15T14:38:17 *** troutwine has left #aichallenge 2010-09-15T14:38:25 but the code is less that perfect 2010-09-15T14:38:26 I can click pretty fast. click could mean 'have everybody send 50% of forces to this spot' or something to speed things up 2010-09-15T14:38:36 *** McLeopold has joined #aichallenge 2010-09-15T14:38:54 eburnette, thats the problem you need to click *and* control how many 2010-09-15T14:38:58 ooo yours allows me to watch the game? Was about to get greghaynes up and running 2010-09-15T14:39:00 http://www.benzedrine.cx/planetwars/canvas?game_id=1284572451|deepblue|McLeopold.5 2010-09-15T14:39:09 use keys to set the default percentage 2010-09-15T14:39:18 deltor the game looks like it just sends half 2010-09-15T14:39:26 say, pick ten keys, make them 10, 20, etc. 2010-09-15T14:39:26 delt0r, galcon and clones let you set a default percentage 2010-09-15T14:39:29 another come from behind, this time before 200 moves 2010-09-15T14:39:35 Zannick, good idea ---1-9 as 10% 2010-09-15T14:39:50 and 0 as 100% 2010-09-15T14:40:02 + and - as maximum and bare minimum 2010-09-15T14:40:06 what the deal with copyright on all this... galcon --is not really a "free" game is it 2010-09-15T14:40:07 ragehuman 2010-09-15T14:40:24 no, but the contest is free advertising 2010-09-15T14:40:31 Could do as all key combos... 2010-09-15T14:41:02 will I be able to get a rating as a human? 2010-09-15T14:41:19 What is the game advertising? 2010-09-15T14:41:21 but a free "clone" that lets you play as a human? ...don't want to be the one to rock that boat 2010-09-15T14:41:37 Err, what is the contest advertising? 2010-09-15T14:41:40 eburnette, On the TCP server yes 2010-09-15T14:41:47 justin_pdx: galcon, the game it's based on 2010-09-15T14:41:57 justin_pdx, The game galcon 2010-09-15T14:42:03 lol 2010-09-15T14:42:05 as in, how many people have bought the game because of the contest? 2010-09-15T14:42:05 I like the canvas javascript viewer, is yours based on that delt0r? 2010-09-15T14:42:09 delt0r: its only playing against a bot - not a big deal - and no mission or scorring 2010-09-15T14:42:21 Well, Galcon looks like a generic version of the same game I've seen dozens of times on Kongregate. Was Galcon the first? 2010-09-15T14:42:24 eburnette, No --i wrote it from the game spec only 2010-09-15T14:42:41 But its like that... numbers --not ships for fleets 2010-09-15T14:43:06 justin_pdx, doubt it... i know some board games with simliar rulz 2010-09-15T14:43:12 vn game - http://www.benzedrine.cx/planetwars/canvas?game_id=1284574844|Cyndre|bartwe.7.6.3 2010-09-15T14:43:26 and there is a star trek like game thats also quite similar 2010-09-15T14:44:12 I like the way the bartwe bots do those chains, very pretty 2010-09-15T14:46:18 Can you download the source for bots you want to play against? Like, could I get bartwe's bot or is that part of the contest, that your code can be a secret if you want it to be. 2010-09-15T14:47:49 i don't think it works that way heh 2010-09-15T14:49:01 thats what galcon reminds me of ... 2010-09-15T14:49:08 Netrek 2010-09-15T14:49:12 how about jar files without the source, would that be any safer? 2010-09-15T14:49:13 but simpler 2010-09-15T14:49:33 Netrek was fun back in the day 2010-09-15T14:50:11 not all bots compile to jar files 2010-09-15T14:50:44 justin_pdx, the best way to play other bots outside the offical board is the tcp server 2010-09-15T14:51:24 Playing them would be cool. But looking at their source would be cooler. 2010-09-15T14:51:58 justin_pdx, yea--I will be making my source more or less available... 2010-09-15T14:52:14 but i don't expect to win... and I have stronger constraints than the rules 2010-09-15T14:52:33 first of all there is no way i want to be taking a whole second 2010-09-15T14:52:43 delt0r: 2010-09-15T14:52:55 delt0r: What do you mean, stronger constraints than the rules? 2010-09-15T14:53:04 Well i am writing a RTS 2010-09-15T14:53:10 I am up to the AI part 2010-09-15T14:53:12 is there like a open repository of bots by other people, i.e. bots that are smarter than the example bots? 2010-09-15T14:53:18 this is a good way to test ideas 2010-09-15T14:53:38 But i can't burn the cpu like a bot thats built to win 2010-09-15T14:53:57 *** bix0r has quit IRC (Quit: Page closed) 2010-09-15T14:54:15 the RTS game has 1000 of units and about 2-5 "game turns" per second 2010-09-15T14:54:37 *** genericbob has joined #aichallenge 2010-09-15T14:54:58 oh so you are going to use the AI for this contest in your RTS? 2010-09-15T14:55:53 *** ct_ has joined #aichallenge 2010-09-15T14:56:10 jmcarthur, pretty much 2010-09-15T14:56:19 langweilig 2010-09-15T14:56:20 ;) 2010-09-15T14:56:20 Anyone else here think that the spec should state a limit for the number of planets? 2010-09-15T14:56:20 clearly there will be differences/ 2010-09-15T14:56:47 mightybyte, it should be either 42 or "heap" 2010-09-15T14:57:10 Argh! I hate it when my bot just stops for no reason and no error 2010-09-15T15:00:01 eburnette: The biggest no-error I found was that my planets were trying to issue orders that would send more ships from that planet per turn than they had. That causes you to forfeit. 2010-09-15T15:00:29 justin_pdx, But usually with a message? no 2010-09-15T15:01:02 I am using java and when I run the commands to test my bot against a test bot the game runner doesn't give any output except that the enemy wins. 2010-09-15T15:01:13 the biggest error i am finding is trying to send negative ships from a planet 2010-09-15T15:01:35 i believe this arises from having negative ships on a planet, but i don't know how that happened 2010-09-15T15:02:10 I changed the IssueOrder function to test for bad parameters and print an error 2010-09-15T15:02:25 justin_pdx, Well once i release my player... you be pleased to note that it will redirect the err stream and give you message... like "we know who you are hax0r --cheater" 2010-09-15T15:02:38 thats code for an illegal fleet command 2010-09-15T15:02:47 but of course I got the compare backwards, doh 2010-09-15T15:03:12 assert is your friend 2010-09-15T15:03:34 mightybyte: not IMO. does it really affect you that drastically? 2010-09-15T15:03:34 *** McLeopold has quit IRC (Quit: Leaving) 2010-09-15T15:03:58 assert will kill your bot and cause it to timeout and be dc'ed 2010-09-15T15:04:28 mightybyte: i am open to being convinced 2010-09-15T15:04:52 jmcarthur, the more planets... the more the game is about expansion 2010-09-15T15:04:57 a maximum number of planets and maximum distance between planets would constrain the problem a bit 2010-09-15T15:04:59 and only expansion 2010-09-15T15:05:15 so i can see a point 2010-09-15T15:05:21 the point 2010-09-15T15:05:21 i mean it seems to be feasible to do everything in O(N^2) now but if it's not in the future then there's a problem 2010-09-15T15:05:30 i think i will use assert for debugging 2010-09-15T15:05:33 but a turn limit would be better first 2010-09-15T15:05:40 yeah i can see how it changes the value of differing goals (destory opponent vs. expand a lot) 2010-09-15T15:05:44 jmcarthur: I think the running time for most people's approaches is most likely to be O(numPlanets ^ p) 2010-09-15T15:06:49 if we were to decide to constrain the number of planets then we open another avenue for bikeshedding. who wants to start? :) 2010-09-15T15:07:07 If the number of planets in the final contest increased by an order of magnitude or more, it would be a completely different kind of contest. 2010-09-15T15:07:13 bikeshedding? 2010-09-15T15:07:28 delt0r: everybody agrees to build a bikeshed but not on what color it should be, so nothing gets done 2010-09-15T15:07:51 jmcarthur: Nah, it should be a dictatorial decision. 2010-09-15T15:07:58 oh --you mean like committees 2010-09-15T15:08:03 delt0r: lol yeah 2010-09-15T15:08:22 I don't particularly care what the exact number is. I just think it should be stated somewhere. 2010-09-15T15:08:27 well first set the number of turn... 200 -400 seems good 2010-09-15T15:08:40 turns is far more important to me, yeah 2010-09-15T15:08:43 then you can't have big maps cus you won't be able to get there 2010-09-15T15:09:13 and the server is already loaded so i don't think its practical to make the maps much bigger in size or planet count 2010-09-15T15:09:23 delt0r: Turns limits map size, but not map density. 2010-09-15T15:09:34 mightybyte: to an extent they do 2010-09-15T15:09:48 dense maps sux like money maps in starcraft 2010-09-15T15:10:13 honestly i wish we had a few man-made maps with weird properties 2010-09-15T15:10:27 and everything is always at least one turn away 2010-09-15T15:10:50 jmcarthur, yea --you could have some funky ones 2010-09-15T15:11:03 There's a post somewhere in the forums that says the number of planets will be around the same as the current maps, but I think that should be stated prominently in the spec. 2010-09-15T15:11:59 mightybyte, I really wouldn't worry about it. if you have something thats going to work on the 100 maps you have, the chances are that its going to just as good as everyone else with different maps 2010-09-15T15:12:14 i wrote the current spec, but as a contestant i can't make a dictatorial decision. we need an organizer to do it 2010-09-15T15:12:51 delt0r: ...unless I'm using a O(n^3) algorithm which works fine now and they crank n up to several hundred for the contest. 2010-09-15T15:13:10 so if one of you guys can get an organizer to declare a number for these things, i will make a patch for the spec page 2010-09-15T15:13:20 mightybyte, and that going to hit *all of us* the same 2010-09-15T15:13:36 anyway i bet real money its not going to change that much 2010-09-15T15:13:37 delt0r: i see mightybyte's point. it affects his choice of algorithm 2010-09-15T15:13:42 delt0r: True, but it creates a disincentive to make more complex algorithms now. 2010-09-15T15:13:42 *** num1_ has quit IRC (Ping timeout: 245 seconds) 2010-09-15T15:13:54 same as for turn count 2010-09-15T15:13:57 In my mind it essentially penalizes "better" algorithms. 2010-09-15T15:14:05 well, i dunno about "better" algorithms 2010-09-15T15:14:12 Thet's why I used quotes. 2010-09-15T15:14:15 heh 2010-09-15T15:14:27 it penalizes a certain class of potentially better algorithms 2010-09-15T15:14:59 I still think your safe assuming the maps you have are representative 2010-09-15T15:15:14 Has anybody noticed any limitations regarding submitting orders? I have a bot that keeps submitting fleet orders which seems to be ignored when I get the game state back (fleet is gone). 2010-09-15T15:15:22 It penalizes a class of algorithms that can reasonably considered to have a good chance of having a higher upper bound of skill. :) 2010-09-15T15:15:27 I would have like more diversity personally... but hay-- 2010-09-15T15:15:47 delt0r: You know what they say about the word "assume"... 2010-09-15T15:15:53 oh yea 2010-09-15T15:16:20 but its not really practical to change it that much... 2010-09-15T15:16:40 Getting enough games in a day for good ranking is already a problem 2010-09-15T15:16:44 *** sinan has joined #aichallenge 2010-09-15T15:16:55 Frankly I think the contest could also be interesting with maps having thousands of planets. 2010-09-15T15:17:04 *** tapwater has joined #aichallenge 2010-09-15T15:17:12 *** num1_ has joined #aichallenge 2010-09-15T15:17:23 kjeldahl: if the trip length is 1 then the fleet will have already landed by the time you get the state back 2010-09-15T15:17:24 me too--but your going to need thousands of turns for that 2010-09-15T15:17:40 But it would be a totally different focus than one constrained to <100 2010-09-15T15:17:53 jmcarthur, not so many maps have planets that close --it turns out 2010-09-15T15:18:02 delt0r: yeah, but some do 2010-09-15T15:18:13 I wish none did 2010-09-15T15:18:13 jmcarthur: Thanks, but it doesn't register in the ships on the planet either... 2010-09-15T15:18:15 delt0r: i was suggesting a possible reason for the behavior that kjeldahl is seeing 2010-09-15T15:18:23 right 2010-09-15T15:18:25 kjeldahl: ah. what engine is this? 2010-09-15T15:18:41 *** ct_ has quit IRC (Quit: Page closed) 2010-09-15T15:19:01 jmcarthur: The modified ones, that output errors to stderr, labeled PlayGame-1.2.jar 2010-09-15T15:19:04 kjeldahl, are you flushing...but the server must get the "go" 2010-09-15T15:19:12 nevermind 2010-09-15T15:19:28 kjeldahl: maybe ping a1k0n about it. he knows more about that code than i 2010-09-15T15:19:43 *** retybok has joined #aichallenge 2010-09-15T15:19:55 Is there a way to get hold of a big dump of game logs? 2010-09-15T15:20:26 I don't need it yet, but I figure I will need one at some point and I think it could be interesting to other participants 2010-09-15T15:20:26 retybok, from the main site? 2010-09-15T15:20:26 retybok: there is a proposal for a daily tarball 2010-09-15T15:20:36 cool 2010-09-15T15:20:36 jmcarthur: awesome! 2010-09-15T15:20:40 retybok: it is likely to be distributed via bitorrent, iirc 2010-09-15T15:20:43 jmcarthur: i've never seen anyone who used darcs except for haskell programmers, i have no idea why this is 2010-09-15T15:20:49 very cool 2010-09-15T15:20:51 what is the "showGame" format anyway... 2010-09-15T15:21:09 dstufft: why it is that nobody else seems to use it or why it is that it's mostly haskell programmers? 2010-09-15T15:21:10 jmcarthur: Thanks, I'll dig a bit more before doing that. Probably something at my end as usual. Just struggling a bit. 2010-09-15T15:21:22 My gameplayer saves games by almost using the same format as in game 2010-09-15T15:21:28 dstufft: i was first introduced to it by steve dekorte of the Io programing language, for the record 2010-09-15T15:21:39 not :6,78,4,3,2,4,5 ???? 2010-09-15T15:21:55 dstufft: who is not a haskell programmer. and this was before darcs 2, even 2010-09-15T15:22:20 *** retybok has quit IRC (Client Quit) 2010-09-15T15:22:22 why haskell programmers love it so much, i swear in my (obviously unscietific) experience, haskell'ers flock to it like no tommarow lol 2010-09-15T15:22:23 dstufft: but yeah, i am also confused by this. darcs is quite remarkable, IMO 2010-09-15T15:22:35 i tried darcs once 2010-09-15T15:22:37 dstufft: probably haskellers are just more exposed to it 2010-09-15T15:22:41 *** retybok has joined #aichallenge 2010-09-15T15:22:42 long time ago 2010-09-15T15:22:46 Ah, found it. The joys of learning new languages. :-) 2010-09-15T15:22:47 dstufft: i love that i can cherry pick at will 2010-09-15T15:22:54 it was horribly slow and i was like eh 2010-09-15T15:22:57 dstufft: ah things have improved a ton since then on the efficiency front 2010-09-15T15:23:04 I'm a haskell programmer and I prefer git to darcs. 2010-09-15T15:23:04 i hear they fixed that though 2010-09-15T15:23:14 ^5 git buddy 2010-09-15T15:23:37 I switched from Hg to git just for github >.> 2010-09-15T15:23:48 and since then found tons of shit i don't know how i lived without 2010-09-15T15:24:06 i used darcs a bit, hated it, switched to git, loved it, switched back to darcs to try it out after hearing it was sped up, fell in love with it, and now hate git 2010-09-15T15:24:22 github is the main thing i miss 2010-09-15T15:24:30 github is awesome 2010-09-15T15:24:43 patch-tag just can't compare 2010-09-15T15:24:53 jmcarthur: Have you tried darcsden? 2010-09-15T15:25:45 *** odinsbane has quit IRC (Quit: out) 2010-09-15T15:26:10 jmcarthur: though to be fair, patch theory seemed overlly obtuse to me, so I probably didn't "get" darcs 2010-09-15T15:26:47 the git parable helped me immensely at understanding git 2010-09-15T15:27:18 dstufft: it's actually a lot simpler than most of the literature lets on 2010-09-15T15:27:19 *** Palmik has quit IRC (Remote host closed the connection) 2010-09-15T15:28:00 i just thought it was funny, i never see anyone using darcs except for haskell programmers lol 2010-09-15T15:28:05 dstufft: basically just think of it as a set of patches rather than a tree of history, and understand that some patches depend on others (but only in terms of content rather than in terms of time), and you've got it 2010-09-15T15:28:33 patch theory is not at all necessary to understand darcs 2010-09-15T15:28:55 won't you end up with a ton of interdependent patches in any codebase of signifcant history 2010-09-15T15:29:23 dstufft: a few, but most such dependencies only make sense anyway 2010-09-15T15:29:55 dstufft: for the most part, unrelated patches will not depend on each other 2010-09-15T15:30:22 it's different for the likes of git, since all commits are strictly dependent on all commits before them 2010-09-15T15:30:29 you can reorder, but then you change identity 2010-09-15T15:30:47 *** amstan has quit IRC (Ping timeout: 245 seconds) 2010-09-15T15:32:23 *** Accoun has quit IRC () 2010-09-15T15:32:52 mightybyte: i haven't tried darcsden, no 2010-09-15T15:32:56 on an unrelated note 2010-09-15T15:33:07 I really need to see if i can compile error emails and such sending 2010-09-15T15:33:24 *** hellman_ has quit IRC (Remote host closed the connection) 2010-09-15T15:33:33 i'm just completly unmotivated to use svn d: 2010-09-15T15:33:35 D: * 2010-09-15T15:34:05 yeah i hate svn 2010-09-15T15:34:23 *** amstan has joined #aichallenge 2010-09-15T15:34:23 *** ChanServ sets mode: +o amstan 2010-09-15T15:34:54 it's just so fucking slow 2010-09-15T15:35:24 that and i never remember the stupid commands 2010-09-15T15:35:26 for branching and shit 2010-09-15T15:35:46 *** phreeza has joined #aichallenge 2010-09-15T15:35:50 and now to make my bot strong 2010-09-15T15:35:52 *** phreeza has quit IRC (Client Quit) 2010-09-15T15:35:59 jmcarthur: http://darcsden.com/ 2010-09-15T15:36:27 also, CIFS YOU FRUSTRATE ME I WISH INOTIFY WORKED 2010-09-15T15:37:48 mightybyte: yeah i've seen that page, but there is nothing about features or anything there. i have to register just to see what's offered 2010-09-15T15:38:16 *** amstan has quit IRC (Client Quit) 2010-09-15T15:38:40 *** amstan has joined #aichallenge 2010-09-15T15:38:40 *** ChanServ sets mode: +o amstan 2010-09-15T15:40:14 it seems most vcs's have a site that is a (usually poor) imitation of github 2010-09-15T15:42:40 dstufft: i agree, github's the best 2010-09-15T15:43:53 bitbucket is good enough 2010-09-15T15:44:45 *** maiden has joined #aichallenge 2010-09-15T15:47:01 nothing compares to github 2010-09-15T15:50:38 *** phreeza has joined #aichallenge 2010-09-15T15:51:27 *** num1_ has quit IRC (Ping timeout: 276 seconds) 2010-09-15T15:51:49 *** pvdbrand has joined #aichallenge 2010-09-15T15:52:05 *** evgenius has joined #aichallenge 2010-09-15T15:57:56 *** Accoun has joined #aichallenge 2010-09-15T16:00:38 *** HakanD has joined #aichallenge 2010-09-15T16:04:27 hello. is everything OK with submissions? do they work? 2010-09-15T16:07:20 no, plenty of things compare with github 2010-09-15T16:07:24 just unfavorably for them 2010-09-15T16:07:46 Well now you're comparing github and oranges. 2010-09-15T16:08:08 *** num1_ has joined #aichallenge 2010-09-15T16:08:36 well, an orange is not a very good repo 2010-09-15T16:09:19 Apples are much better. I keep code in my macintosh all the time. 2010-09-15T16:11:15 ha, found a bug in the .y series 2010-09-15T16:11:19 .7 series 2010-09-15T16:11:27 finally found why .6 was stronger 2010-09-15T16:11:42 that should make for some intresting changes 2010-09-15T16:11:55 bartwe: we should prolly watch out, hehe 2010-09-15T16:12:08 bartwe: i never won against your AI :( 2010-09-15T16:12:13 *** sinan is now known as staiii 2010-09-15T16:12:29 which is your bot ? 2010-09-15T16:12:36 bartwe: staiii 2010-09-15T16:12:38 *** maiden has quit IRC (Quit: Page closed) 2010-09-15T16:12:46 bartwe: i came across yours a couple of time on the tcpserver :) 2010-09-15T16:12:51 justin_pdx: uh, haha. 2010-09-15T16:13:01 bartwe: mine is pretty stupid so far, so no need for any attention, hehe 2010-09-15T16:13:03 staiii: yeah i'm mostly on the tcpserver 2010-09-15T16:16:49 how do you guys continually run it? bash script, cron job, or manually? 2010-09-15T16:17:29 while true; do ./tcpbot ; done 2010-09-15T16:17:51 you might want a random sleep in there to have better random matchings i guess 2010-09-15T16:19:21 max turns is 300? 2010-09-15T16:20:11 *** s_denchev has joined #aichallenge 2010-09-15T16:24:28 *** pvdbrand has quit IRC (Quit: Leaving) 2010-09-15T16:26:51 and when you lost to me bartwe was that when you kmnew your bot had a bug? hehe 2010-09-15T16:30:00 .7 series had a bug, you could trigger it by having 0 units at a planet, sometimes the ownership was mis predicted due to uninitialized values 2010-09-15T16:30:26 which may caused it to under or over defend planets 2010-09-15T16:33:20 *** amstan has quit IRC (Ping timeout: 255 seconds) 2010-09-15T16:33:25 speaking of which, all other ships have crashed and LRO remains in control of the Moon 2010-09-15T16:33:39 lro ? 2010-09-15T16:33:54 lunar reconnaissance orbiter 2010-09-15T16:34:02 har :) 2010-09-15T16:34:22 today is the last day of our exporation mission, and we transition to science mission :) 2010-09-15T16:34:33 too bad the moon is a zero growth location 2010-09-15T16:34:39 no kidding 2010-09-15T16:34:51 you worked on that project ? 2010-09-15T16:35:05 still working on :) 2010-09-15T16:35:17 geh, why does everyone have cool jobs :) 2010-09-15T16:35:31 it actually just occurred to me that my job and this contest had some relation 2010-09-15T16:35:31 I dont, feel better 2010-09-15T16:35:49 bartwe: I have worked at a lot of uncool jobs 2010-09-15T16:36:01 *** bartwe has left #aichallenge 2010-09-15T16:36:28 *** bartwe has joined #aichallenge 2010-09-15T16:36:31 woops 2010-09-15T16:36:53 bartwe disconnect, so I win! 2010-09-15T16:37:00 grrr 2010-09-15T16:37:10 *** pgpaskar_ has joined #aichallenge 2010-09-15T16:38:12 lmao 2010-09-15T16:38:38 woot I am ranked 186 now 2010-09-15T16:40:35 *** ermau|work has quit IRC (Quit: ermau|work) 2010-09-15T16:41:06 *** Vi0 has quit IRC (Ping timeout: 264 seconds) 2010-09-15T16:41:08 w00t, my submission of the default python package is ranked 689. 2010-09-15T16:42:47 your contest entry skills are pretty good! 2010-09-15T16:45:09 *** damag has quit IRC (Quit: Leaving) 2010-09-15T16:45:11 It's like getting the "name:___" question right on the SAT. 2010-09-15T16:45:54 ooh, i got that one right 2010-09-15T16:48:22 *** num1_ has quit IRC (Ping timeout: 265 seconds) 2010-09-15T16:49:33 *** staiii has quit IRC (Ping timeout: 245 seconds) 2010-09-15T16:51:11 *** Kastra has quit IRC (Ping timeout: 240 seconds) 2010-09-15T16:53:29 *** boegel has quit IRC (Quit: Leaving) 2010-09-15T16:55:49 woot top 10 finally 2010-09-15T16:56:13 does somebody have a script to test on all maps against the local bots? 2010-09-15T17:00:51 TapewormGalactos: there are lots floating around... search the forums 2010-09-15T17:00:57 there are some on the fourms 2010-09-15T17:01:01 forums even 2010-09-15T17:01:28 will do 2010-09-15T17:02:02 *** zaph has joined #aichallenge 2010-09-15T17:02:53 *** braddunbar has quit IRC (Ping timeout: 245 seconds) 2010-09-15T17:07:46 *** zaph has quit IRC (Quit: Bye) 2010-09-15T17:17:17 *** zaph has joined #aichallenge 2010-09-15T17:17:26 *** amstan has joined #aichallenge 2010-09-15T17:17:26 *** ChanServ sets mode: +o amstan 2010-09-15T17:18:42 Hi all, how is it going ? 2010-09-15T17:20:46 awesome 2010-09-15T17:21:24 I am on holiday away from internet (apart from phone) so I cannot upload or TCP for a few more days. It's killing me - apart from the sunshine and water 2010-09-15T17:21:26 *** dshreyas has joined #aichallenge 2010-09-15T17:21:51 lot's of time to lay in the sun and think about optimal implementations :) 2010-09-15T17:22:30 Hey, has anybody here tried to setup the ai-contest server themselves? 2010-09-15T17:22:32 and I'm sure you can compile your bot and the tcp client for android and have it run on your phone :D 2010-09-15T17:23:04 I'm trying to set one up locally and kinda stuck on somethings 2010-09-15T17:23:16 deepblue> good play 2010-09-15T17:23:25 zaph: get a server 2010-09-15T17:23:26 That would be the nail in the coffin of everyone's productivity, ai-contest on a phone, no way of escaping it 2010-09-15T17:23:31 I did find the bug that costs me a few games when I play second - noticed it was playing differently 2010-09-15T17:23:52 *** mightybyte has quit IRC (Ping timeout: 265 seconds) 2010-09-15T17:24:07 Hey amstan, I guess just the guy I was looking for. Wonder if you could help me with the setup? 2010-09-15T17:24:42 *** grimm_ has joined #aichallenge 2010-09-15T17:24:49 The web site says 'upload blocked' when viewed from phone - seems short sighted :-) 2010-09-15T17:25:02 I just had a couple of questions like is there any documentation regarding the setup? 2010-09-15T17:25:36 zaph: playing second ? 2010-09-15T17:25:40 *** joeb_ has joined #aichallenge 2010-09-15T17:25:45 *** num1_ has joined #aichallenge 2010-09-15T17:25:49 like how do we create the database (tables, columns...etc) and where're the other stuff located (the daemon for running the game?) 2010-09-15T17:26:01 dshreyas: setup what? 2010-09-15T17:26:07 oh.. 2010-09-15T17:26:13 contestbot: sampledb 2010-09-15T17:26:14 amstan: sampledb = http://csclub.uwaterloo.ca/~amstan/aichallenge/sample.sql 2010-09-15T17:26:17 install that 2010-09-15T17:26:21 Q:can you store information between turns in a game, or is each turn everything reset 2010-09-15T17:26:26 no 2010-09-15T17:26:31 wait 2010-09-15T17:26:31 oooh neat! 2010-09-15T17:26:35 Thanks man! 2010-09-15T17:26:35 you can, yes. 2010-09-15T17:26:37 dshreyas: then... idk.. just download the repo 2010-09-15T17:26:37 yes you can between turns 2010-09-15T17:26:45 not between matches 2010-09-15T17:26:45 but you will always get all the information every turn 2010-09-15T17:26:49 oh ok cool 2010-09-15T17:26:58 dshreyas: that sql contains all users, and only like 5 games 2010-09-15T17:27:01 your program runs 1 match 2010-09-15T17:27:09 sweet, that helps a lot! 2010-09-15T17:27:13 dshreyas: the games take most space, so most of them were removed 2010-09-15T17:27:14 so everything in memory keep turns information 2010-09-15T17:27:25 Edcba: if I was player 2 I had hardcore my starting planet as 1 , not that straightforward but that was the effect. 2010-09-15T17:27:38 that should be fine! 2010-09-15T17:27:38 ok lol 2010-09-15T17:27:53 *** evgenius has left #aichallenge 2010-09-15T17:27:58 *** smhanov__ has joined #aichallenge 2010-09-15T17:28:03 hey guys 2010-09-15T17:28:05 i didn't see how you could make a bug when there is no real 'color' in that game 2010-09-15T17:28:11 ie nobody is first 2010-09-15T17:28:15 is there anything i can do to help to get ruby support going? 2010-09-15T17:28:19 contestbot: help 2010-09-15T17:28:19 dshreyas: (help [] []) -- This command gives a useful description of what does. is only necessary if the command is in more than one plugin. 2010-09-15T17:28:21 if you're player 2, your starting planet has id 2 2010-09-15T17:28:38 even though you think you have id 1 2010-09-15T17:29:13 yes i understood Zannick but i wouldn't ever think about messing with planet ids anyway 2010-09-15T17:29:31 Edcba: I save the starting planets on first turn, but forgot to make it static, so it was reverting to default on turn 2 2010-09-15T17:29:37 *** nigg_ has joined #aichallenge 2010-09-15T17:29:45 grimm_: make a starter kit if there isn't already one, other then that not much really 2010-09-15T17:29:53 ok makes more sense 2010-09-15T17:29:59 so, there is one 2010-09-15T17:30:03 grimm_: the main holdup on new languages, is us (the admins) trying to stablize the game server 2010-09-15T17:30:04 nice bug then 2010-09-15T17:30:18 and it looks like your compile_anything script already supports it 2010-09-15T17:30:33 and fix the bugs that keep creeping up and rendering the game system unusable 2010-09-15T17:30:39 but there's no row for the language in the db... 2010-09-15T17:30:40 mmm 2010-09-15T17:30:40 Edcba: very subtle effects but I was noticing it 2010-09-15T17:30:53 *** nigg_ has quit IRC (Client Quit) 2010-09-15T17:30:54 *** smhanov_ has quit IRC (Ping timeout: 276 seconds) 2010-09-15T17:31:00 grimm_: basically if there's a starter kit, and the compile_anything supports it, all that _we_ need to do is install any runtimes etc 2010-09-15T17:31:04 and add it to the db 2010-09-15T17:31:07 and test it 2010-09-15T17:31:12 i see 2010-09-15T17:31:17 it's just a matter of limited amount of man hours 2010-09-15T17:31:28 *** Bobng has quit IRC (Ping timeout: 240 seconds) 2010-09-15T17:31:31 i can imagine 2010-09-15T17:31:42 and greater severity issues 2010-09-15T17:31:45 *** sigh has joined #aichallenge 2010-09-15T17:31:58 *** McLeopold has joined #aichallenge 2010-09-15T17:32:10 like just today i fixed a bug that was preventing anything from compiling (edge case that caused the whole compile chain to blow up and get stuck) 2010-09-15T17:32:30 so what are the current stability problems you guys are facing? we do a lot of high-availability/scalability bits at work, maybe I can help? 2010-09-15T17:32:37 ah 2010-09-15T17:32:40 grimm_: mostly issues with the coding 2010-09-15T17:32:48 ah 2010-09-15T17:32:57 it wasn't written in the greatest way 2010-09-15T17:33:05 the code is a bit ... convoluted yes :/ 2010-09-15T17:33:06 it's like a house of cards 2010-09-15T17:33:12 lol 2010-09-15T17:33:20 something goes wrong in one part, and things some crashing down 2010-09-15T17:33:28 yah 2010-09-15T17:33:30 we've been working on getting a lot of things fixed though 2010-09-15T17:33:37 *** madsy has quit IRC (Quit: leaving) 2010-09-15T17:33:46 cool 2010-09-15T17:33:59 danielvf just fixed the replay database so that we store the replays in their own table, and compresses 2010-09-15T17:34:02 so, is this the same system you used for last year? 2010-09-15T17:34:11 took our db down from like 3gbs to 11mbs lol 2010-09-15T17:34:19 yikes 2010-09-15T17:34:23 er 3gb and 11mb 2010-09-15T17:34:39 grimm_: i don't know i wasn't around last year 2010-09-15T17:34:45 ah 2010-09-15T17:35:14 i just volunteered to help the other day when the server kept blowing up 2010-09-15T17:35:24 Is the new match schedule running ? I notice I have played a few top-20 games recently 2010-09-15T17:35:29 i see 2010-09-15T17:35:37 and became an admin, same with danielvf 2010-09-15T17:36:04 and have you been able to make some headway at least? 2010-09-15T17:36:11 zaph: i do believe that we have a new scheduler in place, it's still not perfect imo 2010-09-15T17:36:13 i was helping out a lot at the beginning of the month with minor patches 2010-09-15T17:36:34 grimm_: yes, at the moment the server is relatively stable, a lot of the big issues are now taken care of 2010-09-15T17:36:41 awesome 2010-09-15T17:36:42 but then my time grew shorter and i determined i would not be able to contribute much more as an admin 2010-09-15T17:36:43 grimm_: now it's mostly edge cases that keep poping up 2010-09-15T17:36:50 i'm quite curious thoug 2010-09-15T17:36:54 benzedrine.cx is switching game data file formats, i'll have to shut down the server for a couple of minutes, and previous game data will be lost, sorry about that. 2010-09-15T17:37:07 are you running all the games in qemu instances? 2010-09-15T17:37:38 Dstfft: yeah, at least it is an improvement over random 2010-09-15T17:37:51 zaph: i think currently we select a random player, and then we generate a random number, 5, 10, 50, 1000 etc, and pick someone who is that number away from the random person 2010-09-15T17:39:07 personally i think we should pick player 1 by first getting a random player with the fewest games played, until every current submission has atleast N games (maybe 10? or 5?), and then player 2 is selected by what will increase the rankings confidence (since the ranking tries to estimate your ELO score, it's not a true ELO) 2010-09-15T17:39:18 *** smhanov__ has quit IRC (Ping timeout: 240 seconds) 2010-09-15T17:39:26 and verify that two players are not the same, and havn't played a game against each other in the last Y games 2010-09-15T17:39:53 oh, and once every player has atleast X games, then pick the submission with the oldest last played valu 2010-09-15T17:40:07 that sounds like the patch danielvf made 2010-09-15T17:40:27 the patch i saw from danielvf, unless he's made another one 2010-09-15T17:40:27 dstufft ok, not the one I thought they were going with, but a bit better. I think your idea is ok but needs to also look for players with no games in x minutes to prevent starving 2010-09-15T17:40:43 my patch got rejected in favor of that one 2010-09-15T17:40:48 just selcts a random player, and then picks someone within X distance of them 2010-09-15T17:40:51 which is okay 2010-09-15T17:40:54 where X is a random number 2010-09-15T17:41:05 People can keep submitting to jump to front of queue which is bad 2010-09-15T17:41:23 zaph: that's true, i didnt think about that 2010-09-15T17:41:36 adding that feature isn't as hard now, i believe 2010-09-15T17:41:46 changing the scheduler isn't hard really 2010-09-15T17:42:23 *** HakanD has quit IRC (Ping timeout: 240 seconds) 2010-09-15T17:42:26 the issue is just picking a good way to determine player1 and player 2 2010-09-15T17:42:48 If nobody has done it by Monday then I will take a look when I get back 2010-09-15T17:42:56 coinflip? 2010-09-15T17:43:04 my method was just pick player 1 based on recent games 2010-09-15T17:43:12 *** Pitel has quit IRC (Quit: KTHXBYE) 2010-09-15T17:43:22 then pick player 2 based on distance (and recent games if possible) 2010-09-15T17:43:25 ok, back up 2010-09-15T17:43:36 in a way that balances new submissions getting a good number of games asap, all submissions getting a fair chance for playing games, and making the ranking system more confident 2010-09-15T17:44:26 *** deepblue has quit IRC (Ping timeout: 252 seconds) 2010-09-15T17:44:29 Zannick problem with that is I just keep submitting and going to 0 games. 2010-09-15T17:44:54 But it is close to correct 2010-09-15T17:45:02 the scheduler will pick more than one game per run 2010-09-15T17:45:16 well the other issue is, it's nice to give more then one game to a brand new submission 2010-09-15T17:45:24 over someone who might have had 100 games already 2010-09-15T17:45:27 so even if you do that, you will only be at 0 games until your first game that run 2010-09-15T17:45:41 Yep, just needs balance 2010-09-15T17:45:55 and it won't completely starve people unless they get really unlucky with that 2010-09-15T17:46:05 there's a lot of balance needed 2010-09-15T17:46:09 this is a tricky problem 2010-09-15T17:46:10 http://ai-contest.com/forum/viewtopic.php?f=19&t=568#p3533 2010-09-15T17:46:14 yea 2010-09-15T17:47:48 *** tobiassjosten has joined #aichallenge 2010-09-15T17:48:05 the other thing is, we can obviously use a different schedular for the finals 2010-09-15T17:48:08 *** dhartmei has quit IRC (Quit: good night) 2010-09-15T17:48:20 so this one can include stuff for "quality of life" issues 2010-09-15T17:48:34 If player 1 is based off games, player 2 could be based off time + elo and it solves both issues 2010-09-15T17:48:59 afk to take daughter to doctor ! 2010-09-15T17:50:15 When uploading c-sharp submissions, you just put them straight into a zip file (with main file named MyBot.cs ofc), togeather with Fleet.cs, Planet.cs and PlanetWars.cs, or do you need to have them in special folders or anything? 2010-09-15T17:50:41 they all have to be in the same folder 2010-09-15T17:50:51 no other restrictions 2010-09-15T17:50:57 Ok thanks :) 2010-09-15T17:50:58 er 2010-09-15T17:51:14 they all have to be in the same folder, and there has to be a source file "MyBot.cs" (or MyBot.) 2010-09-15T17:51:41 yes, thank you very much :) 2010-09-15T17:52:09 *** Utkarsh has quit IRC () 2010-09-15T17:53:40 Oh, and the note about "if you dont get the email wheter or not your compile was successful within 5 minutes, reupload it", should i reupload or is it just taking its time becouse theres alot of submissions going on? 2010-09-15T17:53:45 *** zaph has quit IRC (Quit: Bye) 2010-09-15T17:55:23 *** num1_ has quit IRC (Ping timeout: 245 seconds) 2010-09-15T17:57:18 *** sigh has quit IRC (Remote host closed the connection) 2010-09-15T17:58:01 yay, upto 105 2010-09-15T17:58:51 Go, CyndreBot, go! 2010-09-15T17:59:24 *** sugarshark has quit IRC (Quit: ERC Version 5.3 (IRC client for Emacs)) 2010-09-15T17:59:25 *** TapewormGalactos has quit IRC (Ping timeout: 265 seconds) 2010-09-15T18:00:13 contestbot: topic replace 2 Channel Logs: http://contestbot.hypertriangle.com/ 2010-09-15T18:00:14 *** contestbot changes topic to "Official Google AI Challenge: http://ai-contest.com/ || Channel Logs: http://contestbot.hypertriangle.com/ || No stupid questions. Don't ask to ask. || Developers wanted, contact j3camero. || Sorry about the slowness. || Latest Idea, decentralization: http://tinyurl.com/decentralizedai" 2010-09-15T18:00:14 aww 2010-09-15T18:00:32 *** smhanov has joined #aichallenge 2010-09-15T18:03:08 Anyone else using ulope's Python kit? I'm having problems with it not reporting my own fleets with p.attacking_fleets or p.reinforcement_fleets. 2010-09-15T18:04:47 *** smhanov_ has joined #aichallenge 2010-09-15T18:05:36 *** AJC has joined #aichallenge 2010-09-15T18:05:56 are the bots in the leaderboard getting better or are the scores only increasing because of the evaluations? 2010-09-15T18:06:04 *** s_denchev has quit IRC (Ping timeout: 252 seconds) 2010-09-15T18:07:00 yeah are they 2010-09-15T18:07:11 *** smhanov has quit IRC (Ping timeout: 240 seconds) 2010-09-15T18:07:17 *** grimm_ has quit IRC (Quit: Page closed) 2010-09-15T18:08:29 *** mightybyte has joined #aichallenge 2010-09-15T18:10:46 tobiassjosten: did you try looking at the code? 2010-09-15T18:11:27 AJC: i'm sure they are getting better 2010-09-15T18:11:51 amstan: Yes, and I can't find anything. Though I'm pretty green with Python so I might as well miss something obvious. 2010-09-15T18:11:52 *** mightyby1e has joined #aichallenge 2010-09-15T18:15:06 http://www.benzedrine.cx/planetwars/canvas?game_id=152 2010-09-15T18:15:33 A game with "phases" 2010-09-15T18:17:43 cool game. blue lost sending that 74 ship probably. 2010-09-15T18:17:51 right at the start... fleet I mean. 2010-09-15T18:18:18 Yea, I didn't see that at first. 2010-09-15T18:18:31 Sometimes that works out. deepblue does it often. 2010-09-15T18:18:58 *** mightyby1e has quit IRC (Quit: leaving) 2010-09-15T18:21:14 *** Appleman1234 has joined #aichallenge 2010-09-15T18:21:40 *** Any_Key has quit IRC (Ping timeout: 252 seconds) 2010-09-15T18:22:33 McLeopold: I'll have to take a look! 2010-09-15T18:22:52 *** Any_Key has joined #aichallenge 2010-09-15T18:23:00 McLeopold: the map seems a bit too big to pull that off. 2010-09-15T18:26:07 Way too big for that trick. 2010-09-15T18:26:56 http://www.ai-contest.com/visualizer.php?game_id=4480015 2010-09-15T18:27:05 wonder how to break that deadlock 2010-09-15T18:27:32 either side could go for their 26 planet nearby, others seem like too big of an investment. 2010-09-15T18:28:55 I'm surprised red didn't try and hold the center. 2010-09-15T18:29:34 yeah, he would have taken over again presumably had the game continued. 2010-09-15T18:29:35 blue was stocking a lot of ships in the lower left 2010-09-15T18:29:56 needs a better distance metric to fix that 2010-09-15T18:30:36 Zannick: what do you mean by distance metric? 2010-09-15T18:31:08 it's better to have ships closer to other planets than on the ones in the corners 2010-09-15T18:31:32 http://ai-contest.com/forum/viewtopic.php?f=18&t=621 2010-09-15T18:31:51 turn 18, he launches most of his fleets towards his corner planet 2010-09-15T18:32:00 yeah, somewhat paranoid. 2010-09-15T18:32:09 http://ai-contest.com/forum/viewtopic.php?f=18&t=621 perl starter package 2010-09-15T18:32:41 Is the TCP server down? 2010-09-15T18:32:45 For bot testing. 2010-09-15T18:34:46 a turn around! http://www.ai-contest.com/visualizer.php?game_id=4485350 2010-09-15T18:35:07 wow 2010-09-15T18:37:12 blue stockpiling again, bottom right. 2010-09-15T18:37:14 *** braddunbar has joined #aichallenge 2010-09-15T18:38:05 *** amstan has quit IRC (Ping timeout: 265 seconds) 2010-09-15T18:39:41 *** zerd_ has quit IRC (Ping timeout: 240 seconds) 2010-09-15T18:40:57 *** zerd has joined #aichallenge 2010-09-15T18:41:32 how many of you would say you are using more AI techniques than you are manually tweaking parameters? 2010-09-15T18:41:54 Uhm, the visualizer at benzedrine.cx/planetwars isn't working. Is that just me or is it broken? 2010-09-15T18:42:05 *** Caffeinated has joined #aichallenge 2010-09-15T18:42:15 AJC: i've moved onto less parameter tweaking 2010-09-15T18:42:18 *** braddunbar has quit IRC (Ping timeout: 264 seconds) 2010-09-15T18:42:26 Mornings 2010-09-15T18:42:26 gonna take a few days to have my next versions done i think 2010-09-15T18:42:47 nevermind I see it works in Firefox but now with Chromium 2010-09-15T18:42:58 antimatroid: does the AI technique you're using have a name? 2010-09-15T18:43:43 i'm intending to use a few things 2010-09-15T18:43:56 and intending to keep them private mostly 2010-09-15T18:44:02 hehe 2010-09-15T18:44:09 which parts are public? :-) 2010-09-15T18:44:27 not sure, prety much what everyone else is doing, index future information as best you can 2010-09-15T18:44:40 but that's not really ai 2010-09-15T18:44:53 index? 2010-09-15T18:45:05 expected states :P 2010-09-15T18:45:10 ah 2010-09-15T18:45:19 no, not really :-) 2010-09-15T18:46:03 but it's a matter of making the most of the information you have, so make it accessible then come up with ideas to use that to own the other player 2010-09-15T18:46:48 antimatroid, good to hear, I've been working pretty much solely on organizing information 2010-09-15T18:47:03 I tried doing some planning doing future state estimation at the same time and it's just way too slow to work. 2010-09-15T18:47:18 that's what i've mostly been doing, i think i've almost finished that, just cleaning up and working on the next stages of implementation :) 2010-09-15T18:47:21 also ruled out monte-carlo tree search, the space is too big! 2010-09-15T18:47:30 AJC: i rewrote planet wars 2010-09-15T18:47:41 mine still runs in under like 0.01 seconds 2010-09-15T18:47:44 and uses 1.2mb memory 2010-09-15T18:47:56 antimatroid: how much planet max ? 2010-09-15T18:48:17 they haven't specified one, but unless they want a lot of bots to fall over in the final tournament, it wont be much more than 23 2010-09-15T18:48:31 antimatroid: your prediction of the whole game in 0.01s? how far ahead? 2010-09-15T18:48:39 all known fleets 2010-09-15T18:48:49 i doubt there is 'far' for him ;) 2010-09-15T18:49:53 Quick q - most likely "bending" the rules, but is there anything against sending lots of fleets with just 1 attack in, to take advantage of the bots who process every fleet info? (Seems as relevant a time as any =P) 2010-09-15T18:50:22 of course no 2010-09-15T18:50:27 Caffeinated: I don't think it's against the rules. 2010-09-15T18:50:42 but if you send the exact same fleets at the exact same place, then it's unsportsman :-) 2010-09-15T18:50:42 but i doubt it is a winning strategy 2010-09-15T18:50:55 edcba: it'll help against the non C++ bots ;-) 2010-09-15T18:51:07 depends on how they are coded 2010-09-15T18:51:09 Does PlayGame merge fleets to/from the same place on the same turn? 2010-09-15T18:51:15 not yet obviously 2010-09-15T18:51:21 the TCP server does however 2010-09-15T18:51:32 well, if the logic is there anyways, you're just sending it in bigger amounts of data 2010-09-15T18:51:35 Caffeinated: then you'd have to work around that :-) 2010-09-15T18:51:54 you wont hurt me anymore than yourself by spamming fleets :P 2010-09-15T18:52:00 i've set it up to account for that 2010-09-15T18:52:21 antimatroid: how so? 2010-09-15T18:52:23 Wasn't planning to do it.. It just occured to me after a few of the bots timed out ;) 2010-09-15T18:52:50 Caffeinated: A bot just has to merge them as a direct counter 2010-09-15T18:52:51 AJC: i'll leave that one as an exercise to the reader 2010-09-15T18:52:52 i think sending at same place is "wrong" 2010-09-15T18:53:09 else of course there is no problem sending 1 fleet at each planet 2010-09-15T18:53:14 antimatroid: you didn't say how far ahead you were predicting, beyond the last fleet arrival. 2010-09-15T18:53:39 *** Queue29 has joined #aichallenge 2010-09-15T18:53:40 AJC: max distance for each planet, has that many future states 2010-09-15T18:53:47 there might be memory issues 2010-09-15T18:54:22 the engine on benzedrine seems to send malformed state when there are many fleets 2010-09-15T18:54:35 antimatroid: you mean your code regroups fleets going to the same place. you still have the issue of 1-fleets sent to different places on purpose. 2010-09-15T18:54:35 hehe 2010-09-15T18:56:12 And you'd have to adjust for different travel times, sent over a number of turns may not be the same, unless you're only caring about destination.. 2010-09-15T18:56:29 you don't care where a fleet is sent from, only its arrival time, you can merge any meeting that ;) 2010-09-15T18:56:42 could still end up with lots of fleets, but so will everyone else 2010-09-15T18:56:51 *** Mathnerd314 has joined #aichallenge 2010-09-15T18:57:05 are you using C++? 2010-09-15T18:57:08 i am 2010-09-15T18:57:29 *** hornairs has joined #aichallenge 2010-09-15T18:57:43 yeah, I think C++ should win this contest :-) 2010-09-15T18:57:48 Good ole Java here ;-) 2010-09-15T18:57:53 not if C takes it :) 2010-09-15T18:57:59 I'm on Python until I figure everything out. 2010-09-15T18:58:00 *** num1_ has joined #aichallenge 2010-09-15T18:58:11 It's actually refreshing to use it, since my work insists on TCL >__> 2010-09-15T18:58:14 *** farbe has joined #aichallenge 2010-09-15T18:59:47 I'm hoping they have a ruby package soon. 2010-09-15T19:03:12 *** syntaxglitch has joined #aichallenge 2010-09-15T19:03:25 I don't get the current obsession with RoR.... 2010-09-15T19:04:15 1/2 the job vacancies I see atm are for Ruby in one flavour or another... 2010-09-15T19:04:19 Me either 2010-09-15T19:07:10 RoR was a hype 5 years ago, no it's just one of many frameworks 2010-09-15T19:07:18 now 2010-09-15T19:07:26 *** sigh has joined #aichallenge 2010-09-15T19:08:07 *** kozlovsky has joined #aichallenge 2010-09-15T19:13:11 oh god, why do i leave assignments so late, your mums a quotient space of a vectory space by your sisters subspace 2010-09-15T19:14:19 *** inverselimit has joined #aichallenge 2010-09-15T19:14:25 *** jaspervdj has quit IRC (Quit: leaving) 2010-09-15T19:14:43 are you swearing now? 2010-09-15T19:15:16 *** phreeza has quit IRC (Quit: Computer has gone to sleep) 2010-09-15T19:16:19 *** tobiassjosten has quit IRC (Quit: Leaving) 2010-09-15T19:16:56 *** jacob_strauss has joined #aichallenge 2010-09-15T19:18:26 *** amstan has joined #aichallenge 2010-09-15T19:18:26 *** ChanServ sets mode: +o amstan 2010-09-15T19:19:33 *** AJC has left #aichallenge 2010-09-15T19:21:21 *** caytchen has quit IRC () 2010-09-15T19:21:43 hello 2010-09-15T19:22:22 hullo 2010-09-15T19:22:29 *** McLeopold has quit IRC (Quit: Leaving) 2010-09-15T19:23:48 *** RainCT has quit IRC (Remote host closed the connection) 2010-09-15T19:26:32 *** kozlovsky has quit IRC (Quit: Page closed) 2010-09-15T19:29:43 hmm.. I lose one match and drop 40 places.. Win the next and gain.... -10... wtf? 2010-09-15T19:30:27 And the win was higher ranked than me O_o 2010-09-15T19:30:55 tcp? rating is updated every 7 minutes 2010-09-15T19:31:15 *** ath888 has joined #aichallenge 2010-09-15T19:31:50 Which is the tcp one? I'm on about the actual ai-contest.com site? 2010-09-15T19:33:08 ah 2010-09-15T19:33:16 http://www.benzedrine.cx/planetwars 2010-09-15T19:37:01 *** num1_ has quit IRC (Ping timeout: 276 seconds) 2010-09-15T19:37:30 Caffeinated: uncertainty of score is a hidden factor 2010-09-15T19:37:47 Im at 108 on ai-contest.com 2010-09-15T19:38:00 Caffeinated: i'm hoping that my proposed ranking system will be approved and go into effect (show and sort on lower confidence bound rather than on mean) 2010-09-15T19:38:13 the algorithm is the same, it's just a different displayed value 2010-09-15T19:38:17 *** delt0r___ has joined #aichallenge 2010-09-15T19:39:39 I love how the rules are still up for debate XD 2010-09-15T19:39:39 *** delt0r_ has quit IRC (Ping timeout: 255 seconds) 2010-09-15T19:39:51 *** jacob_strauss has quit IRC (Quit: Page closed) 2010-09-15T19:41:33 Caffeinated: if the people that you lost against became much stupider on the next rankings, you could go down too 2010-09-15T19:42:14 oh, is it a ranking of smartness? 2010-09-15T19:42:36 it can't be a ranking of smartness because true smartness would involve memory dumps and unfair tactics ;) 2010-09-15T19:45:16 ShadeS: just... no 2010-09-15T19:45:24 Naktibalda: yes... see ELO on wikipedia 2010-09-15T19:45:42 oh, I know ELO 2010-09-15T19:46:00 I fought that it's for chess, not for smartness :) 2010-09-15T19:46:03 this is like ELO, but with confidence intervals 2010-09-15T19:46:08 my bot wins way more often on the ai-contest.com 2010-09-15T19:46:20 than? 2010-09-15T19:46:40 http://www.benzedrine.cx/planetwars/getratings 2010-09-15T19:46:41 ai-contest has lots of default bots... I stop watching after about 5 turns if they've only sent out one attack ;-) 2010-09-15T19:47:23 whereas if you can be bothered to set up the tcp thing, you're likely taking it a little more seriously 2010-09-15T19:47:44 I want to break into the top 100 2010-09-15T19:48:14 i hit 300 earlier.. pretty impressed with it as a first attempt 2010-09-15T19:48:39 I'm not ready to submit my bot yet 2010-09-15T19:48:45 i was in the top ten briefly 2010-09-15T19:48:47 ...during the beta 2010-09-15T19:48:48 I'm having trouble getting it to do what i want 2010-09-15T19:49:10 can you guys use functions in the code or does it all have to be nested ifs and/or for/while loops? 2010-09-15T19:49:17 http://ai-contest.com/visualizer.php?game_id=4487917 I love how my bot kinda flows (not sure how to explain it) 2010-09-15T19:49:31 ShadeS I'm having that problem, mine's really good in the middle stages, but starts off badly and could kill off matches a lot quicker than it does 2010-09-15T19:49:54 And you can do whatever you want as long as it takes less than a second per turn 2010-09-15T19:50:50 yeah I don't exactly know how to check for that 2010-09-15T19:50:56 I have a simple solid gameplan 2010-09-15T19:51:19 caffieinated: to select planets I scan all I dont own, find out cost to take them and distance away - I then use the max distance I see as my gauge and all planets are ranked as shipCost = shipCost - ((maxDist - dist) * growthRate) or something close to that 2010-09-15T19:51:19 but it's weird because it's a lmost recursive in the way it works 2010-09-15T19:52:24 Cyndre: Seems a fairly solid plan, though I found that sending out my own growthrate left me open.. 2010-09-15T19:53:02 sending out growth rate is useless, unless you send it to your own planet 2010-09-15T19:53:06 Caffeinated: thats just the ranking part to decide which planet to take 2010-09-15T19:53:20 Cyndre I know, I was watching the link you gave 2010-09-15T19:53:29 Natibalda: not true - as soon as my bot grows faster I send growth rate to enemy planet 2010-09-15T19:53:50 Caffeinated: ahh, only once my bot is ahead in growth rate - and I use it absorb your growth rate 2010-09-15T19:54:27 Caffeinated: if you grow 20 a turn and Im at 22 a turn Ill immediatly start attacking your planet 2010-09-15T19:55:03 Cyndre: Hadn't considered that, I'm currently finding the "best" planet based on distance, growthrate and such, then send out what I can afford to not have locally given current fleet positions 2010-09-15T19:55:40 works fairly well, except for the maps with smaller planets around, where it decides "That planet directly opposite me is growing nicely.. I'll take it" 2010-09-15T19:55:41 Caffeinated: one thing I need to add is fleet evaluation on my "am I ahead code" 2010-09-15T19:56:33 cool 2010-09-15T19:57:09 Caffeinated: my bots worst thing is when the enemy attacks planets across the map - my bot prematurely starts attacking you while your fleet is in flight 2010-09-15T19:57:42 and over reinforcing 2010-09-15T19:58:12 Cyndre: Mine developed reinforcing from nowhere.. Not a clue how either, since I'm only scanning through NotMyPlanets for targets... 2010-09-15T19:58:43 http://ai-contest.com/visualizer.php?game_id=4489369 that was a very wierd bug 2010-09-15T19:58:59 *** Queue29 has quit IRC (Remote host closed the connection) 2010-09-15T19:59:24 Caffeinated: I did the exact same thing - I added fleet scanning and replaced the current planetID variable with the destination planetID by accident 2010-09-15T20:00:14 so planetID = planet.PlanetID() and then in the same function lower down - planetID = fleet.DestinationPlanet().PlanetID() 2010-09-15T20:00:28 Cyndre: LOL - love the tiny planet on the opposite side 2010-09-15T20:00:43 not sure what my bot was doing - was actually messed up 2010-09-15T20:00:59 couldnt kill an 11 sized planet with 250 ships 2010-09-15T20:01:38 Might work better sending out target.growthrate + 1 for when the growths are vaguely even? 2010-09-15T20:02:04 if someone is around that built the visualizer - could you add a graph of ships vs ships and growthRate vs growthRate 2010-09-15T20:02:19 left it with 11 ships though 2010-09-15T20:02:22 Or add something so if your planet is above 70 for example you just take your target out 2010-09-15T20:02:41 not sure why it was sending like that 2010-09-15T20:03:38 think I need to do a dist + 1 when I caluclate taking enemy planets 2010-09-15T20:04:51 there is already a ships vs ships graph 2010-09-15T20:05:04 not on the client visualizer, on the canvas one 2010-09-15T20:05:25 the one on ai-contest right? 2010-09-15T20:05:31 yes 2010-09-15T20:05:38 *** Heckel has joined #aichallenge 2010-09-15T20:05:52 look under the buttons 2010-09-15T20:06:05 should add a growth rate chart as well - maybe a bar graph in behind or a ---- line for it 2010-09-15T20:06:25 *** Heckel has quit IRC (Remote host closed the connection) 2010-09-15T20:06:33 *** Any_Key has quit IRC (Ping timeout: 265 seconds) 2010-09-15T20:06:57 then you can see that even though ships were even the enemy gained growth faster then you 2010-09-15T20:07:15 *** Yoshi-TS4 has joined #aichallenge 2010-09-15T20:08:34 Damnit.. I hate typing half a line of code and losing my attention span.... Come back to it to find "if (...." and not a clue what I was going to do... 2010-09-15T20:10:17 I find ctrl z helps 2010-09-15T20:10:31 undo 10 steps, then slowly redo them and it will get you to remember 2010-09-15T20:10:59 Not in vim it doesn't :P *ducks out before inevitable vi/emacs/butterfly arguement* 2010-09-15T20:11:32 Im using idle - not an argument I would waste time on other then trying to convince you that it really doesnt matter 2010-09-15T20:11:34 u in vim 2010-09-15T20:12:17 heh, I know, but I'd also only just reopened.. u a few times took me back to "Already at oldest change" 2010-09-15T20:12:42 :( 2010-09-15T20:12:56 Couldn't of been that important I guess.. 2010-09-15T20:13:22 its good if you never close it - I get distracted by a pin dropping sometimes and cant remember if I was debuging a problem or trying something new 2010-09-15T20:15:48 Caffeinated: on the 100 maps - I get 93 wins on dualbot and 97 on ragebot now - and can beat ragebot on map 30 - should try it 2010-09-15T20:16:24 Cyndre: Rage on 30 beats me in under 50 moves.... 2010-09-15T20:16:47 So disheartening that every change I made just for that map made it quicker.... 2010-09-15T20:16:48 51 moves to win 2010-09-15T20:17:17 *** dshreyas has quit IRC (Quit: Page closed) 2010-09-15T20:18:08 I need to edit the script for testing against the maps - kinda want to see the 3 match's rage bot beats me on 2010-09-15T20:20:22 and only because I dont take more then 1 planet a turn to start 2010-09-15T20:20:58 maybe thats what I will do - take 1 planet the first turn - then determine % of ships in air and base future decisions on that 2010-09-15T20:21:21 First turn is the hardest to get right IMO 2010-09-15T20:22:04 From there you're pretty much sorted, but not knowing anything about the opponent means you have nothing to base it on 2010-09-15T20:22:49 *** rwa has joined #aichallenge 2010-09-15T20:25:00 dualbot (map15, map 4, map7, map86, map9, map96, map99) Ragebot (map22, map54, map94) are the maps I lose on 2010-09-15T20:27:21 map22 is nasty 2010-09-15T20:28:48 *** hornairs has quit IRC (Ping timeout: 240 seconds) 2010-09-15T20:30:13 *** dshreyas has joined #aichallenge 2010-09-15T20:31:26 Beat 22 in 58 moves vs Rage 2010-09-15T20:31:39 nice 2010-09-15T20:32:02 just add a +2 to my ships required calculation 2010-09-15T20:32:09 Went with your method of sending based on growth, but *2 2010-09-15T20:32:31 ? 2010-09-15T20:32:51 oh, you mean sending the growthrate of a planet at it? That was a bug for sure 2010-09-15T20:33:20 *** ath888 has quit IRC (Quit: Page closed) 2010-09-15T20:33:20 Heh, I liked it :P Does a decent job of stalling their ship growth so they can't attack 2010-09-15T20:33:36 what I meant was my bot compares my growthRate and the enemys growthRate - as soon as my bot is ahead it will send all ships to 1 target enemy planet 2010-09-15T20:33:39 focus fire 2010-09-15T20:33:48 Obviously, with checks to stop a 250 planet falling to an 11 ;-) 2010-09-15T20:34:11 not even sure wtf was going on there 2010-09-15T20:35:08 *** DirtyKeyboard has joined #aichallenge 2010-09-15T20:35:18 if I worked on just the issues of my bot I could probably get top 50 2010-09-15T20:38:27 I'm starting my quantum multi-verse bot to compete with what I see going on here 2010-09-15T20:39:51 just founda bug with my bot - it would modify for inbound enemy fleets - then modify for inbound friendly fleets, and each one would call the planet.NumShips() - fuk 2010-09-15T20:40:22 boo 2010-09-15T20:40:46 *** pp__ has joined #aichallenge 2010-09-15T20:41:27 no prize? 2010-09-15T20:41:28 no one believed that kid that kept calling wolf 2010-09-15T20:41:32 prize plz 2010-09-15T20:41:39 for? 2010-09-15T20:41:42 hmm.. Inbound friendlies.... Interesting concept 2010-09-15T20:41:48 cash prize for contest 2010-09-15T20:41:50 boo 2010-09-15T20:42:03 joeb_ resume highlight will bring you a cash bonanza 2010-09-15T20:42:10 no need for explicit money prize 2010-09-15T20:42:28 although certainly it would be nice 2010-09-15T20:42:37 exactly, heh 2010-09-15T20:42:44 but i don't see why not 2010-09-15T20:42:50 googles cash strapped? 2010-09-15T20:43:24 It's only using Google's name, they have nothing to do with it... And why should they pay you to enter a competition they're making nothing from? 2010-09-15T20:43:49 as a way to find good candidates who can make them even more money 2010-09-15T20:44:04 At that point they'll give you money ;-) 2010-09-15T20:44:52 google has nothing to do with it, then why let the contest use their name? 2010-09-15T20:44:53 Few of my job interviews involved "code this" before an actual interview.. Something google's size it makes sense to make it competitive 2010-09-15T20:45:21 joeb_ publicity... coz clearly not enough CS students had heard of them.... 2010-09-15T20:45:28 joeb_ would you rather win the google ai contest planet wars or ai-contest.com ai contest planet wars? 2010-09-15T20:45:46 if there's no $$ prize, both equal :) 2010-09-15T20:46:19 you don't think headhunters arnt watching the top competitors? 2010-09-15T20:46:55 a1k0n got a promotion to a better job from winning last year (and many job offers) 2010-09-15T20:47:13 *** retybok has quit IRC (Ping timeout: 276 seconds) 2010-09-15T20:47:14 if it didnt say google that wouldnt have happened 2010-09-15T20:47:17 so it's just for people who want a better job? 2010-09-15T20:47:28 no, its for publicity 2010-09-15T20:47:38 I know for a fact that where I work is monitoring it.. Helps that we have an office league going ;-) 2010-09-15T20:47:47 oooo very cool 2010-09-15T20:48:23 if you work with machine learning should use it to hand out bonus's 2010-09-15T20:48:26 *** Yoshi-TS4 has quit IRC (Quit: Minus One IRC) 2010-09-15T20:49:31 *** Appleman1234 has quit IRC (Quit: Leaving) 2010-09-15T21:00:32 *** Janzert1 is now known as Janzert 2010-09-15T21:01:08 *** danielvf has joined #aichallenge 2010-09-15T21:03:18 *** danielvf_ has joined #aichallenge 2010-09-15T21:03:18 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-15T21:03:18 *** danielvf_ is now known as danielvf 2010-09-15T21:03:53 i take 23 place last year, no any good hapened with me 2010-09-15T21:04:35 Its a talking point on a CV too though 2010-09-15T21:05:46 Guessing majority of entrants are students? Makes a lot more sense for them to have more than and have something a little different too 2010-09-15T21:06:00 hey anything to differiate yourself 2010-09-15T21:06:05 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-15T21:06:27 *** danielvf has joined #aichallenge 2010-09-15T21:06:37 Yeah. It's also a lot of just plain fun. 2010-09-15T21:06:53 many worck 2010-09-15T21:06:58 The rest is just justifying participation. :)P 2010-09-15T21:07:14 And many, should be, working. 2010-09-15T21:13:18 anyone know how to just pipe stderr to a file so I don't need logging in the code? 2010-09-15T21:13:47 *** vulture- has joined #aichallenge 2010-09-15T21:14:05 You could wrap your bots script in another script 2010-09-15T21:14:13 shell script or batch script 2010-09-15T21:16:13 add '2>&1 1>/dev/null >some_file.log' to the end of a command, might be excessive, but my bash-fu isn't as good as it should be :P 2010-09-15T21:16:31 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-15T21:17:18 *** danielvf has joined #aichallenge 2010-09-15T21:23:05 if anyone else has rusty Bash-fu, the same syntax (less the /dev/null block device) applies to Tae Kwon DOS as well, incidentally 2010-09-15T21:23:15 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-15T21:23:44 Caffeinated: doesnt' seem to work for me using the Java play game. hmm 2010-09-15T21:26:33 *** danielvf has joined #aichallenge 2010-09-15T21:26:50 inverselimit: Odd, Just tried it and it seems to be working for me 2010-09-15T21:27:48 http://ai-contest.com/forum/viewtopic.php?f=18&t=621 perl starter package 2010-09-15T21:28:17 * Cyndre hurrys up and waits for his newbot to play some game 2010-09-15T21:28:18 s 2010-09-15T21:29:09 *** num1_ has joined #aichallenge 2010-09-15T21:29:09 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-15T21:29:23 *** danielvf has joined #aichallenge 2010-09-15T21:34:15 yay - http://ai-contest.com/visualizer.php?game_id=4492912 2010-09-15T21:39:29 *** Any_Key has joined #aichallenge 2010-09-15T21:40:33 *** Snowplan has joined #aichallenge 2010-09-15T21:40:59 Is there a problem with Distance(planet,planet) 2010-09-15T21:42:52 Snowplan: if there is that would explain a bug I can't seem to find in my code 2010-09-15T21:42:54 not in the python one 2010-09-15T21:44:03 int(ceil(sqrt(dx * dx + dy * dy))) that is correct isnt it? 2010-09-15T21:44:06 anyone know if are the maps all symmetrical (to be fair), or will there ever be lopsided maps ? 2010-09-15T21:45:12 *** dmj111 has quit IRC (Quit: ERC Version 5.3 (IRC client for Emacs)) 2010-09-15T21:46:52 vulture - Pretty sure they all are symmetrical 2010-09-15T21:47:15 it's unspecified, but probably maps will be designed to be fair 2010-09-15T21:47:43 yeah only asking cuz I couldnt find it specified 2010-09-15T21:47:52 vulture-: either point or line symmetric 2010-09-15T21:48:20 should that be in the spec? 2010-09-15T21:48:41 jmcarthur: idk though.. 2010-09-15T21:48:45 *should it be added to the spec, i mean 2010-09-15T21:49:04 if we do introduce an unbalanced map, it should still even out 2010-09-15T21:49:17 would there be enough game to do that? 2010-09-15T21:49:20 i'm kind of doubtful 2010-09-15T21:49:24 *games 2010-09-15T21:49:24 I've done some AI contests in the past where the host decided to just use randomly generated boards, and the winner usually ended up being whoever had the random initial state advantage, whereas symmetry removed that 2010-09-15T21:49:33 I am all for non asymetric maps 2010-09-15T21:49:42 non-asymmetric :P 2010-09-15T21:49:55 vulture-: 2 player maps? 2010-09-15T21:49:59 yeah 2010-09-15T21:50:02 i think we should just specify that they should be symmetrical 2010-09-15T21:50:19 14 games per minute.... 1000 bots... not gonna be enough time to balence out the non-asymmetricless ones... 2010-09-15T21:50:21 is this contest eventually gonna be multi player 2010-09-15T21:50:22 i can't see any other way to get meaningful results from the contest 2010-09-15T21:50:25 vulture-: and how many games would each ai play? 2010-09-15T21:50:25 on 3+ player FFA AI games I always notice the loser is automatically chosen by whoever gets teamed up on :D 2010-09-15T21:50:27 Snowplan: not this contest 2010-09-15T21:50:37 I dunno, doesnt really matter 2010-09-15T21:50:43 cause it seems to be set up for it 2010-09-15T21:50:58 Snowplan: there is a little untested infrastructure, but no plans to use it for this contest 2010-09-15T21:51:02 vulture-: it does.. if it's few games there's a much higher chance that you'll get fucked by a crappy map 2010-09-15T21:51:05 if you're 2v1 it doesnt really matter how good your AI for the 1 is, the 2 will always beat the 1 in the upper ranks 2010-09-15T21:51:26 vulture-: it's 1v1 all the time anyway.. 2010-09-15T21:51:32 ok thanks 2010-09-15T21:51:33 yeah so it wouldnt matter 2010-09-15T21:51:39 amstan: even if there are many games, if the map advantage determines the game 90% of the time, say, then it would take a *ton* of games to even that out 2010-09-15T21:51:49 jmcarthur: good point 2010-09-15T21:51:51 ok.. 2010-09-15T21:51:52 1v1 + symmetry --> more games played = more accurate ranking of your AI 2010-09-15T21:51:52 well 2010-09-15T21:51:54 put it in 2010-09-15T21:52:05 okay. i'll add it either tonight or tomorrow 2010-09-15T21:52:08 vulture-: symmetry does not result in more games played 2010-09-15T21:52:09 oh I see what you're saying 2010-09-15T21:52:32 you're saying if you play the asymmetric maps an uneven number of times or whatever 2010-09-15T21:52:40 it just needs more games to be accurate 2010-09-15T21:52:47 even if we always played two games between each bot, swapping positions, it wouldn't make up for the asymmetry 2010-09-15T21:52:54 *** num1_ has quit IRC (Ping timeout: 245 seconds) 2010-09-15T21:53:01 nah, you just play two games and swap start positions 2010-09-15T21:53:03 I guess you could do each asymmetric map twice, just flip the initial player positions 2010-09-15T21:53:17 if the asymmetry accounts for a 90% advantage then it pretty much just means that we waste those games on a nearly predetermined outcome 2010-09-15T21:53:20 that'd be fine 2010-09-15T21:53:21 vulture-: no.. if you play 1000 games on uneven maps there's a higher chance to even out the advantages/disadvantages of the map positions 2010-09-15T21:53:30 Dumb ? but if I change my pw.java will it work on the contest server 2010-09-15T21:53:35 vulture-: it it's fewer, it won't be evened out, so it'll suck 2010-09-15T21:53:48 1000 might not even you out at the higher ranks 2010-09-15T21:54:03 the change of 500:500 advantage is pretty small 2010-09-15T21:54:07 yes, when you submit you zip all the files you need and the files in the zip are used 2010-09-15T21:54:10 that is, if the asymmetry gives any X% advantage, then we might as well be just throwing away X% of the game outcomes 2010-09-15T21:54:13 *chance 2010-09-15T21:54:20 ok thats great 2010-09-15T21:54:23 or more even 2010-09-15T21:54:25 jmcarthur: not that much, but you're right 2010-09-15T21:54:26 jmcarthur: not really, play 2 games 2010-09-15T21:54:29 thought maybe it was only mybot 2010-09-15T21:54:32 jmcarthur: fewer 2010-09-15T21:54:40 jmcarthur: and swap start positions 2010-09-15T21:54:40 think about it: 2010-09-15T21:54:44 Cyndre: not sufficient 2010-09-15T21:54:44 anyway.. put it in the specs 2010-09-15T21:54:50 jmcarthur: why not? 2010-09-15T21:54:57 jmcarthur: playing double the games with swapped positions would solve this 2010-09-15T21:55:02 I think you could do it by playing two games swapped 2010-09-15T21:55:11 vulture-: that needs more games 2010-09-15T21:55:17 Cyndre: consider the case that the advantage is 99%. that means we have a 99%*99% chance that the outcome of those two games is just (1,1) 2010-09-15T21:55:20 Cyndre: pretty pointless 2010-09-15T21:55:21 I mean for EACH asymmetric map 2010-09-15T21:55:26 vulture-: ^^ 2010-09-15T21:55:41 jmcarthur: no.. 2010-09-15T21:55:47 # games played overall irrelevant (more is better obviously), but for an asymmetric map, play twice with positions swapped 2010-09-15T21:55:48 why not? 2010-09-15T21:55:58 jmcarthur: totally changes the dynamics and brings more then just best search tree 2010-09-15T21:55:59 jmcarthur: you get 90% advantage the first game, then you swap the second game and gain -90% advantage 2010-09-15T21:56:01 it evens out 2010-09-15T21:56:09 amstan: they don't just subtract like that 2010-09-15T21:56:15 each game is an independent sample 2010-09-15T21:56:26 How do you output from your bot in java? Can you put messages in the log? 2010-09-15T21:56:27 jmcarthur: not if you use the same map and the same bot 2010-09-15T21:56:33 not really - if you play the 2 bots on the same bot over and over its identical results 2010-09-15T21:56:40 er same map over 2010-09-15T21:56:47 justin_pdx: stderr, use the modified game engine of a1k0n's design 2010-09-15T21:56:49 if in the first game you have a 90% chance of winning and in the second a 10% chance, it doesn't even out to a 50% chance overall 2010-09-15T21:56:50 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-15T21:57:01 no, that is not how probability works 2010-09-15T21:57:05 jmcarthur: but it gives both players identical odds 2010-09-15T21:57:06 Where do I find the modified game engine? 2010-09-15T21:57:20 it gives both players identical chances 2010-09-15T21:57:21 Cyndre: that's true, but that's not my point 2010-09-15T21:57:22 *** danielvf has joined #aichallenge 2010-09-15T21:57:23 yes 2010-09-15T21:57:24 justin_pdx: forums, it's stickied in tech problems 2010-09-15T21:57:25 i agree with that 2010-09-15T21:57:27 jmcarthur, it does help if someone wins 2-0 though... 2010-09-15T21:57:29 Thank you. 2010-09-15T21:57:30 *** Any_Key has quit IRC (Read error: Operation timed out) 2010-09-15T21:57:32 but it also throws away information about player skill 2010-09-15T21:57:39 not true 2010-09-15T21:57:43 Caffeinated: yeah, *if*. it makes that much less likely though 2010-09-15T21:57:52 maybe your AI is better at making a come back from a losing situation 2010-09-15T21:58:00 or too wreckless when you're in a better position 2010-09-15T21:58:16 gives more AI possibilities than just symmetry 2010-09-15T21:58:17 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-15T21:58:30 yes 2010-09-15T21:58:31 if it doesnt matter, then the two games will even out 2010-09-15T21:58:33 *** danielvf has joined #aichallenge 2010-09-15T21:58:35 most bots are not like that 2010-09-15T21:58:38 if it makes any difference, maybe it wont be 50:50 2010-09-15T21:58:44 anyone have a link to the goatse map from tron? 2010-09-15T21:58:51 yeah the games even out, but it makes then less useful in determining skill 2010-09-15T21:59:03 actually that makes it more useful 2010-09-15T21:59:16 as I mentioned, [20.52.28] gives more AI possibilities than just symmetry 2010-09-15T21:59:21 when you have a 90%*90% chance of just getting (1,1) then you only have a 100%-90%*90% chance of getting any other result 2010-09-15T21:59:43 what's 90%*90% 2010-09-15T21:59:50 81% 2010-09-15T21:59:53 vulture-: he's kinda right.. 2010-09-15T21:59:56 no I mean whats your number 2010-09-15T22:00:03 what other number? 2010-09-15T22:00:09 jmcarthur: the map generator could be made so the uneveness is lower 2010-09-15T22:00:11 maybe you have a 90% chance of winning a "9:1 odds game", but a 20% chance of winning a "1:9 odds game" 2010-09-15T22:00:11 oh 2010-09-15T22:00:16 because your AIs are different 2010-09-15T22:00:27 vulture-: i'm using a theoretical advantage given to one player over the other in an asymmetrical map 2010-09-15T22:00:43 yeah sure. most AIs aren't going to do necessarily better in one situation than the other though 2010-09-15T22:00:54 we're just decreasing the signal to noise ratio 2010-09-15T22:00:58 thats not necessarily always the case 2010-09-15T22:01:14 *** hornairs has joined #aichallenge 2010-09-15T22:01:14 there might be like one or two games that a single bot plays that gives extra information in this case 2010-09-15T22:01:17 out of all of them 2010-09-15T22:01:19 all AIs are different 2010-09-15T22:01:23 (as a guess) 2010-09-15T22:01:24 its completely plausible 2010-09-15T22:01:33 I doubt that 2010-09-15T22:01:33 plausible is not good enough to justify the noise, IMO 2010-09-15T22:01:36 I've done a lot of AI games 2010-09-15T22:01:43 9:1 probably not 2010-09-15T22:01:48 but 4:6 maybe easily 2010-09-15T22:02:01 jmcarthur: you're right though.. we want less noise 2010-09-15T22:02:19 if you can qualitatively prove that you aren't just talking out of your ass then you might stand a chance of convincing me that it's worth it. got any real-life numbers? 2010-09-15T22:02:29 even 4:6 is not good for us 2010-09-15T22:02:30 but then its a depth search problem, not an ai problem 2010-09-15T22:03:05 jmcarthur: you know how a unbalanced map system could work? 2010-09-15T22:03:09 well, let's see how my latest bot version does 2010-09-15T22:03:25 jmcarthur: have an unbalanced factor associated with each map, then multiply the score of the winner/looser by that 2010-09-15T22:03:32 I fail to see how you've disproved anything here 2010-09-15T22:03:46 vulture-: i haven't attempted to 2010-09-15T22:03:46 more input posibilities = more reflections of how your AI performs 2010-09-15T22:03:52 maybe we should use only one map with 4 planets 2010-09-15T22:03:58 jmcarthur: a bot that won from a really sucky position will get more points than winning from a good position 2010-09-15T22:03:58 it'd be the same right 2010-09-15T22:04:19 vulture-: but i think i've shown fairly well that asymmetrical maps would increase noise. the burden is on you to show that the noise is worth it for any statistical benefits it might have due to wonky AIs 2010-09-15T22:04:31 how does the noise increase exactly 2010-09-15T22:04:33 amstan: how would we come up with some a factor? 2010-09-15T22:04:36 *such a factor 2010-09-15T22:04:40 jmcarthur: previous games 2010-09-15T22:04:52 jmcarthur: win/looses directly 2010-09-15T22:04:53 if asymmetry doesnt make any difference, you end up with 1:1 2010-09-15T22:05:08 if it does, then it means one AI is better than another at dealing with advantage/disadvantage, and you dont end up with 1:1 2010-09-15T22:05:21 vulture-: i've been explaining it already. if we increase the odds that two closely ranked bots end up with the same score overall then we have failed to distinguish them 2010-09-15T22:05:41 and having asymmetric maps would attempt to distinguish them more :P 2010-09-15T22:05:52 it's not about extra bias introduced by the asymmetry. it's about smearing the results around 2010-09-15T22:05:59 vulture-: burden of proof 2010-09-15T22:06:05 no, it's about [20.59.21] if it does, then it means one AI is better than another at dealing with advantage/disadvantage 2010-09-15T22:06:13 vulture-: we still need to play double games then 2010-09-15T22:06:17 yes you do 2010-09-15T22:06:19 vulture-: any numbers for that? 2010-09-15T22:06:19 vulture-: which is not an option at this point 2010-09-15T22:06:22 it still has to be fair/even 2010-09-15T22:06:26 oh, no? 2010-09-15T22:06:34 even double games won't get rid of all the additional noise 2010-09-15T22:06:35 vulture-: our servers are dying as is 2010-09-15T22:06:38 what noise 2010-09-15T22:06:41 haha 2010-09-15T22:06:51 your servers are dying? 2010-09-15T22:06:55 just needs a little caressing optimization 2010-09-15T22:06:56 what di dyou think you were getting indto? 2010-09-15T22:07:16 vulture-: optimizations won't help get more games, it'll only help stability and http response times 2010-09-15T22:07:18 plug in a little home server into the wall and think that'll do it? ;) 2010-09-15T22:07:30 ShadeS: for that we need to get distribution 2010-09-15T22:07:30 You attach the 'google' name to it and wonder why people take notice? 2010-09-15T22:07:34 oh, cpu time isnt your bottleneck? 2010-09-15T22:07:44 vulture-: just take your "more realistic" bias from an asymmetrical map of 4:6. already we are saying that rather than a 25% chance of (1,1) being the result, that's actually a 36% chance 2010-09-15T22:07:46 vulture-: it is.. 2010-09-15T22:07:56 then how doesnt optimization help 2010-09-15T22:07:57 we're losing information if the result is indeed (1,1) 2010-09-15T22:08:00 jmcarthur: again... unfairness coefficients 2010-09-15T22:08:27 amstan: i'm with you there, but we'd need to have a large sample on any given map before applying it to anybody's scores in the first place 2010-09-15T22:08:43 we would still just end up burdening the servers with games that don't even count, if we did it fairly 2010-09-15T22:09:03 jmcarthur: yeah.. the coefficient involves writing more code 2010-09-15T22:09:09 way more 2010-09-15T22:09:32 and running the map through a ton of bots without counting it, unless we are willing to introduce *tons* of unfairness initially 2010-09-15T22:09:56 [21.02.10] we're losing information if the result is indeed (1,1) <-- you dont necessarily lose information if you're using absolutes... you just dont gain any information.. or maybe you can throw out tie games, then you never lose information, but only gain information if the result is (0,2) or (2,0) 2010-09-15T22:10:00 even then, as bots improve we would have to *again* recalculate that factor 2010-09-15T22:10:16 jmcarthur: it could be on the fly 2010-09-15T22:10:26 vulture-: not gaining information from a couple games is losing the potential information from those game in normal play 2010-09-15T22:10:29 so... 2010-09-15T22:10:29 vulture-: we just do not have enough cpu time to do more matches 2010-09-15T22:10:38 optimize! 2010-09-15T22:10:39 from what I just read I'm pretty convinced the only winner is skynet 2010-09-15T22:10:42 amstan: with what initial value? 2010-09-15T22:10:48 jmcarthur: but the information gained from a 2,0 win is worth more then a asymetrical win 2010-09-15T22:10:54 hey, my new bot actually won a match 2010-09-15T22:10:57 I just assumed google had infinite cpu time =/ 2010-09-15T22:10:59 Cyndre: yeah, and the odds of that happening are vastly decreased 2010-09-15T22:11:03 vulture-: you can't optimize! you cannot touch ppl's code 2010-09-15T22:11:11 *** Naktibalda has quit IRC (Remote host closed the connection) 2010-09-15T22:11:11 vulture-: there's one server... 2010-09-15T22:11:14 oh 2010-09-15T22:11:19 jmcarthur: which would eliminate alot of games when the top 100 bots beat the bottom 200 2,0 2010-09-15T22:11:35 Cyndre: i agree that if we could take this to a fixpoint, we'd have about the same amount of working information, but we can't do that 2010-09-15T22:11:50 we have a limited sample size 2010-09-15T22:11:53 jmcarthur: could be 50% for initial 2010-09-15T22:11:55 and we have to make it work 2010-09-15T22:12:06 jmcarthur: nah, just give them a proportianate ranking boost to the odds of it happening and they will skip up the ranks faster eliminating games 2010-09-15T22:12:12 rogue780: gratz :P 2010-09-15T22:12:16 amstan: how many games do you think it would take before that factor goes to about what it should be? 2010-09-15T22:12:32 *** joeb_ has quit IRC (Quit: Page closed) 2010-09-15T22:12:33 Cyndre: that is what amstan is proposing 2010-09-15T22:12:34 jmcarthur: i was thinking.. add 2 columns for each map: player1 wins, player2 wins, then divide them to get coefficient 2010-09-15T22:12:58 amstan: we're talking a lot of work for something that might make a difference of about 1% in rankings 2010-09-15T22:13:03 *** Any_Key has joined #aichallenge 2010-09-15T22:13:04 jmcarthur: it will be inaccurate at first, but after about 10 games i say it's good 2010-09-15T22:13:04 amstan: *ignoring* the noise 2010-09-15T22:13:07 How bout instead of just adding win/lose, adding the number of turns it takes to make a decision? 2010-09-15T22:13:12 I doubt it's a 1% difference 2010-09-15T22:13:31 anyway.. this is not the point 2010-09-15T22:13:36 even with a 90% advantage, if bot A wins in 100 turns and bot B takes 300, A is clearly better 2010-09-15T22:13:41 the point is that we can't implement either method 2010-09-15T22:13:48 jmcarthur: no, what vulture said - throw out tied match's and give elo gain elogain * elogain if 2,0 2010-09-15T22:13:49 Caffeinated: that would require changing the entire ranking system. what we have right now is pretty dang good 2010-09-15T22:14:00 Cyndre: bayeselo doesn't work that way 2010-09-15T22:14:01 running double games requires more cpu power, having coeff requires more coding 2010-09-15T22:14:10 jmcarthur: but you get the general idea? 2010-09-15T22:14:35 look, i see everybody's points, i'm just saying we can't play enough games to justify the extra noise in the samples 2010-09-15T22:14:48 *over a large number of games* this might be better 2010-09-15T22:15:12 well I originally joined this channel thinking asymmetric games could be unfair, but as someone pointed out, you could just run it twice swapping positions, so I thought, hey that'd be fair after all... but if CPU power is limited, seems like thats more of a concern right now :P 2010-09-15T22:15:15 and i think it would have to be more than double the games 2010-09-15T22:15:19 *** justin_pdx has quit IRC (Quit: justin_pdx) 2010-09-15T22:15:21 jmcarthur: No it doesn't, one match is made of 2 "halves", if A wins on both sides, A wins... if B wins on both sides, B wins... if A wins in 100 B wins in 300, A wins.. etc 2010-09-15T22:15:24 jmcarthur: not really, if in tie its broken by turns 2010-09-15T22:15:40 Cyndre: that would require changing the ranking system 2010-09-15T22:15:53 yeah it would 2010-09-15T22:15:55 jmcarthur: how so? this is all before the winner is decided 2010-09-15T22:15:58 Cyndre: or are you proposing that we count each game pair as a single game? 2010-09-15T22:15:58 vulture-: go checkout the repo, then follow the link from the topic for decentralization, you could help with that by coding 2010-09-15T22:16:02 jmcarthur: yes 2010-09-15T22:16:05 Cyndre: which would mean we have to run twice the games 2010-09-15T22:16:08 but the problem statement right now was ambiguous about whether maps were all symmetric or not 2010-09-15T22:16:19 vulture-: unfortunatelly if you want to host a node you won't be allowed to compeete 2010-09-15T22:16:28 jmcarthur: but with what caffeinated suggested we would not lose anything and gain alot 2010-09-15T22:16:38 Cyndre: gain... what exactly? 2010-09-15T22:16:45 a little strangeness in the results? 2010-09-15T22:17:04 guys.. let's stop the argument, this obviously requires more study/experimentation 2010-09-15T22:17:07 does it really mean that a bot is "better" just because it handles itself better in certain situations than others? 2010-09-15T22:17:15 jmcarthur: what would be easier to write a bot for - symetrical maps or asymetrical maps? 2010-09-15T22:17:17 I'm pretty backed up with coding and journal papers as is =/ 2010-09-15T22:17:26 Cyndre: i don't see your point in asking that question 2010-09-15T22:17:40 jmcarthur: thats exactly what it means! 2010-09-15T22:17:43 jmcarthur: Different scenarios than mirrored games. Ability to adapt even with a completely unknown map, rather than knowing that you and the opponent have the same layout 2010-09-15T22:17:47 vulture-: then i just disagree 2010-09-15T22:17:54 look at the board game of Go 2010-09-15T22:17:58 vulture-: it's a different contest if that's the goal 2010-09-15T22:18:10 jmcarthur: is it ai contest or depth search contest? 2010-09-15T22:18:11 you give a handicap of different quantity to one player 2010-09-15T22:18:25 and the relative difference of that handicap is pretty significant in determining how much better your AI is 2010-09-15T22:18:39 vulture-: the goal as i understand it right now is "make my bot better than yours in a 1-on-1 match. the goal that would be necessary for me to solve under your proposal would be "make my bot better than your bot under all sort of biased circumstances" 2010-09-15T22:18:53 i lost a " there 2010-09-15T22:19:00 vulture-: in go the number of pieces put on a board by the end is very close to 50%, in our situation, no 2010-09-15T22:19:04 Cyndre: i still don't see the point 2010-09-15T22:19:08 but its not un biased - your both given the exact same maps 2010-09-15T22:19:10 right which is why my original original first question was if there would ever be biased circumstances 2010-09-15T22:19:26 thats not really an accurate analogy 2010-09-15T22:19:29 jmcarthur: you dont see the point but your saying it changes the entire dynamics to unimaginable poportions 2010-09-15T22:19:36 it does 2010-09-15T22:19:49 jmcarthur: if it changes it that much then there is definatly a point to it 2010-09-15T22:19:53 even the starting player in Go can decide the game 2010-09-15T22:19:58 it means that if you haven't been coding for that goal you need to rewind and rethink 2010-09-15T22:19:59 much less adding a one piece handicap 2010-09-15T22:20:14 vulture-: sure, then what's wrong with just starting fairly? 2010-09-15T22:20:15 yeah I was hoping to have that goal in mind before I started 2010-09-15T22:20:30 whether it was fair or biased would ever occur 2010-09-15T22:20:52 jmcarthur: what were you coding for? (and no Im not trolling, I am trying to understand - symetrical maps or asymetrical would not force me to do anything) 2010-09-15T22:20:53 fair would be simpler; biased could be more interesting 2010-09-15T22:20:57 bias happens naturally in a game. i say it's better if that bias is deserved in the first place than if it's artifically introduced 2010-09-15T22:21:08 Cyndre: neither in particular, at the moment 2010-09-15T22:21:23 Cyndre: but there are tons of strategies for which that would be a major factor 2010-09-15T22:21:36 I think there's a lot more significant amount of work involved in writing an AI to handle biased situation 2010-09-15T22:21:42 Cyndre: for example, if your bot has any concept of early game, mid game, late game, you're automatically screwed in a biased situation 2010-09-15T22:21:49 Cyndre: unless it can also detect that 2010-09-15T22:21:50 so in fact fair would be easier/simpler/less time spent for me 2010-09-15T22:21:57 but I still think biased would be more interesting 2010-09-15T22:22:05 jmcarthur: but wouldnt the point in the game be based on the state of the game not the layout of the board? 2010-09-15T22:22:13 vulture-: perhaps, but "more fair" is another story 2010-09-15T22:22:40 Cyndre: if you know that the board always starts in a biased state then you can code for early game differently than for late game 2010-09-15T22:22:52 Cyndre: to get early advantages, for example 2010-09-15T22:23:03 jmcarthur: the starting planets are matched so you both have same growth 2010-09-15T22:23:03 Cyndre: which, i should add, is a *major* win in this particular game 2010-09-15T22:23:08 what I wouldnt want to ever happen in an AI challenge is where you let the die roll decide which AI is better (e.g., playing biased games, and giving 1 AI more advantage-biased games than another) 2010-09-15T22:23:21 vulture-: yeah that would suck the worst 2010-09-15T22:23:38 I'm all for either fair or biased though, just wanna know which 2010-09-15T22:23:43 Cyndre: starting planets aren't the only thing we care about 2010-09-15T22:23:44 since the rules didnt seem to state it 2010-09-15T22:23:54 vulture-: i'm pretty sure we are going to go with unbiased 2010-09-15T22:24:00 k 2010-09-15T22:24:12 yeah thats like 10 times less coding 2010-09-15T22:24:21 Cyndre: what vulture- said 2010-09-15T22:25:10 since I have virtually no time due to papers/coding that works best for me :D 2010-09-15T22:25:18 heh 2010-09-15T22:25:32 well that was fun. we should argue more later :) 2010-09-15T22:25:38 *** ShadeS has quit IRC (Ping timeout: 240 seconds) 2010-09-15T22:25:42 hah 2010-09-15T22:25:47 not arguing, it was a discussion 2010-09-15T22:25:50 yeah, if only I wasnt sick! 2010-09-15T22:25:57 aw man, get better 2010-09-15T22:25:58 arguing is negative, this was positive 2010-09-15T22:26:01 i hate teh sickness 2010-09-15T22:26:10 there's something going around here idk 2010-09-15T22:26:10 *** Any_Key has quit IRC (Ping timeout: 255 seconds) 2010-09-15T22:26:19 can it spread via irc? 2010-09-15T22:26:23 my oldest is sick to 2010-09-15T22:26:56 until someone makes a multipartite reallife/digital virus, I think you're safe 2010-09-15T22:27:06 awesome 2010-09-15T22:27:40 * jmcarthur envisions strange "Nightfall"-like worlds 2010-09-15T22:27:50 not Nightfall 2010-09-15T22:28:01 damn. right author, wrong story 2010-09-15T22:30:36 I love running a script against all 100 maps for RageBot.. >_> 2010-09-15T22:31:28 i guess The Last Question is close enough, but is still not what i was thinking of. still the right author and subject matter 2010-09-15T22:31:41 (asimov, singularity) 2010-09-15T22:31:46 what are you scorring now caffeinated? and what do you get on that jmcarthur 2010-09-15T22:31:59 38w, 23l 49d.... 2010-09-15T22:32:13 Cyndre: my bot's not ready to run yet 2010-09-15T22:32:16 still crawling 2010-09-15T22:32:20 Might've messed up the weightings.... 2010-09-15T22:32:21 oh 2010-09-15T22:32:38 Im at 94 for Dualbot, and 98 for ragebot now 2010-09-15T22:32:46 but still getting my ass handed to me on tcp server 2010-09-15T22:32:58 39d... can't type 2010-09-15T22:33:35 I dont know why my bot waits for the fleet to land before sending another one 2010-09-15T22:33:53 oh I bet my bot is saying - this is the best planet, send 0 ships 2010-09-15T22:33:55 lmao 2010-09-15T22:34:30 http://www.benzedrine.cx/planetwars/canvas?game_id=1901 2010-09-15T22:36:19 any help on mapping planets? 2010-09-15T22:36:30 *** j3camero has joined #aichallenge 2010-09-15T22:36:51 Wow damn, 98 people in here. That's pretty sweet I would say. 2010-09-15T22:37:03 *** Naith has joined #aichallenge 2010-09-15T22:37:32 my Distance function is not working with me. I'm sure I'll figure it out but rather ask then waste time. 2010-09-15T22:37:37 went from 114 turns to win on map 30 against dualbot to 42 turns once I added and shipCost >= 1 2010-09-15T22:38:17 Snowplan: what language 2010-09-15T22:38:35 java 2010-09-15T22:38:57 Snowplan, how you mean not working? 2010-09-15T22:39:45 hiya jscamero 2010-09-15T22:39:51 *** Any_Key has joined #aichallenge 2010-09-15T22:39:56 j3camero: oh yes.. 2010-09-15T22:40:24 did you notice that we have 1200 registered(in rankings.php) already? 2010-09-15T22:40:25 well Im trying to map by Distance(planet,planet) and It's telling me Distance(int,int) doesn't read argument (planet, planet) 2010-09-15T22:40:34 j3camero: we've had more people than this here :) 2010-09-15T22:40:35 that's 150% of what we had when we ended tron 2010-09-15T22:40:42 which it should or am I wrong 2010-09-15T22:40:43 you want planet.PlanetID() 2010-09-15T22:41:01 *** jmreardon has quit IRC (Quit: jmreardon) 2010-09-15T22:41:03 or the other way around 2010-09-15T22:41:19 ^ 2010-09-15T22:41:27 No, was right first time.. you want the PlanetID() bit 2010-09-15T22:41:54 ok thanks didn't find a PlanetID() 2010-09-15T22:42:02 ill try that 2010-09-15T22:42:07 Actually we have like 2300 people registered. Just some of them are n00bs and haven't figured out how to submit the starter package. We should maybe work on that. 2010-09-15T22:42:22 excellent 2010-09-15T22:42:29 How are people finding the contest? Is it pretty easy to use and such? 2010-09-15T22:42:50 at first laggy. lol 2010-09-15T22:43:04 fun though 2010-09-15T22:43:12 j3Camero: http://65.255.180.155/docmon/ss1.png and then http://65.255.180.155/docmon/ss2.png 2010-09-15T22:43:12 j3camero, easy to get started and hard to get good at... am loving it =) 2010-09-15T22:43:47 Cyndre: we were thinking about that last term 2010-09-15T22:44:06 Cyndre: it is pretty much a development enviroment, it has help, compiling, execution, testing and all that included 2010-09-15T22:44:07 even if its only for the python starter bot 2010-09-15T22:44:08 except an editor 2010-09-15T22:45:11 Cyndre 2010-09-15T22:45:13 ? 2010-09-15T22:45:29 Cyndre: what the shit? Is that real? Or is that a mock-up of a suggested feature? 2010-09-15T22:45:35 mockup 2010-09-15T22:45:39 I wish I had this contest when in college. would have kept me more focused on programing 2010-09-15T22:45:42 you just said an easier way to get started 2010-09-15T22:45:44 :) 2010-09-15T22:45:51 personally i think that feature is out of scope but idk 2010-09-15T22:46:03 glad you like it Snowplan. Tell your friends! 2010-09-15T22:46:24 j3camero: I distributed links in the ponoka hospital 2010-09-15T22:46:48 *** ShaedS has joined #aichallenge 2010-09-15T22:46:50 Have already. One friend is really into it. instead of signing up he's gonna contribute to my bot. 2010-09-15T22:46:52 ponoka albera mental hospital that it is - specifically the ocd department...... j/k but might be funny 2010-09-15T22:47:00 that was an epic failure 2010-09-15T22:47:03 *** dmj111 has joined #aichallenge 2010-09-15T22:48:33 Well I am glad people are liking it. 2010-09-15T22:48:54 could probably borrow some open source paste bin python syntax highlighting to 2010-09-15T22:49:18 I like it when Im not Ranked in the 1000 + 2010-09-15T22:49:22 After we fixed the website and started adding new features, the web metrics are steadily ticking up. pages per visit, bounce rate, time per visit, all slowly grinding in the right direction. I guess people are happy. 2010-09-15T22:49:23 Cyndre: hold on.. 2010-09-15T22:50:10 Cyndre: are you interested in building it? That would be soooo bitchin. I have seen some simple online IDEs before, and I gotta say it's pretty cool. We could build something unprecedented! 2010-09-15T22:50:12 * Cyndre waits for amstan to pull out some multi fauceted html5 ide hes been waiting to demo 2010-09-15T22:50:17 Cyndre: http://www.cdolivet.com/editarea/editarea/exemples/exemple_full.html 2010-09-15T22:50:34 Cyndre: not exactly html5, but it has syntax highlighting 2010-09-15T22:51:06 *** Appleman1234 has joined #aichallenge 2010-09-15T22:51:55 but even just a simple text box that gives them the ability to edit a file - save file to filenameNEW makesure file runs replace bot file or return error of why it wont run/compile 2010-09-15T22:52:40 I have 90% of the other code amstan (have a php script that edits dhcpd.leases, backs it up, checks if it works, and replaces and restarts if fails 2010-09-15T22:52:54 er not leases, but dhcpd.conf 2010-09-15T22:54:04 You can't change directions of fleets, right? 2010-09-15T22:54:18 Cyndre: this is mostly for client side though 2010-09-15T22:54:24 Cyndre: you can use your scripts already 2010-09-15T22:54:36 Cyndre: this is instead of having a