2011-09-03T00:14:08 *** codetiger has joined #aichallenge 2011-09-03T00:14:44 *** amstan has quit IRC (Ping timeout: 245 seconds) 2011-09-03T00:20:05 *** amstan has joined #aichallenge 2011-09-03T00:20:05 *** ChanServ sets mode: +o amstan 2011-09-03T00:41:57 *** codetiger has quit IRC (Remote host closed the connection) 2011-09-03T01:11:24 *** Allied_Envy is now known as AlliedEnvy 2011-09-03T01:19:18 *** ibdknox has joined #aichallenge 2011-09-03T01:21:38 http://pastebin.com/JhWyYgqF 2011-09-03T01:21:42 anyone know why that wouldn't work? 2011-09-03T01:21:59 i want a constructor that takes template objects as parameters :\ 2011-09-03T01:35:50 nevermind 2011-09-03T01:35:56 i had a namespace issue 2011-09-03T01:40:04 *** dvladim has joined #aichallenge 2011-09-03T01:41:18 *** deeper-blue has quit IRC (Quit: deeper-blue) 2011-09-03T01:47:41 *** ibdknox has quit IRC (Remote host closed the connection) 2011-09-03T02:29:29 *** dvladim has quit IRC (Ping timeout: 246 seconds) 2011-09-03T02:33:25 *** dvladim has joined #aichallenge 2011-09-03T02:34:11 *** amstan has quit IRC (Remote host closed the connection) 2011-09-03T02:51:24 *** delt0r_ has quit IRC (Ping timeout: 245 seconds) 2011-09-03T03:04:42 *** delt0r_ has joined #aichallenge 2011-09-03T03:14:31 *** Palmik has joined #aichallenge 2011-09-03T03:19:49 *** dvladim has quit IRC (Ping timeout: 240 seconds) 2011-09-03T03:19:59 *** McLeopold has quit IRC (Read error: Connection reset by peer) 2011-09-03T03:21:56 *** UncleVasya has joined #aichallenge 2011-09-03T03:28:59 *** ibdknox has joined #aichallenge 2011-09-03T03:31:57 *** ibdknox has quit IRC (Remote host closed the connection) 2011-09-03T03:46:07 *** mceier has joined #aichallenge 2011-09-03T03:55:03 *** berak has joined #aichallenge 2011-09-03T04:42:03 *** FireFly has joined #aichallenge 2011-09-03T04:43:33 *** Frantic_ has joined #aichallenge 2011-09-03T04:46:32 *** antimatroid has quit IRC (Ping timeout: 250 seconds) 2011-09-03T04:46:57 * Frantic_ test 2011-09-03T04:48:12 *** Frantic_ has left #aichallenge 2011-09-03T04:48:20 *** antimatroid has joined #aichallenge 2011-09-03T05:31:31 *** boegel has joined #aichallenge 2011-09-03T05:33:16 *** Kingpin13 has joined #aichallenge 2011-09-03T05:36:37 *** mathis has joined #aichallenge 2011-09-03T05:39:25 *** eashoka has joined #aichallenge 2011-09-03T05:39:56 *** boegel has quit IRC (Quit: This computer has gone to sleep) 2011-09-03T06:07:24 *** mathis has quit IRC (Ping timeout: 252 seconds) 2011-09-03T06:19:59 *** Palmik has quit IRC (Remote host closed the connection) 2011-09-03T06:31:24 *** dvladim has joined #aichallenge 2011-09-03T06:37:15 *** Redgis has joined #aichallenge 2011-09-03T06:48:04 *** antimatroid has quit IRC (Ping timeout: 264 seconds) 2011-09-03T06:48:29 *** mathis has joined #aichallenge 2011-09-03T06:49:31 *** antimatroid has joined #aichallenge 2011-09-03T07:03:33 *** antimatroid has quit IRC (Ping timeout: 276 seconds) 2011-09-03T07:12:49 *** antimatroid has joined #aichallenge 2011-09-03T07:46:29 *** dvladim has quit IRC (Ping timeout: 252 seconds) 2011-09-03T08:00:13 *** UncleVasya has quit IRC () 2011-09-03T08:08:03 *** dvladim has joined #aichallenge 2011-09-03T08:26:43 *** dvladim has quit IRC (Read error: Operation timed out) 2011-09-03T08:48:33 *** nux67 has joined #aichallenge 2011-09-03T08:59:53 *** mathis has quit IRC (Disconnected by services) 2011-09-03T09:07:51 *** mathis_ has joined #aichallenge 2011-09-03T09:15:19 *** amstan has joined #aichallenge 2011-09-03T09:15:19 *** ChanServ sets mode: +o amstan 2011-09-03T09:39:44 *** amstan has quit IRC (Ping timeout: 245 seconds) 2011-09-03T10:14:51 antimatroid: it's syntax like that that reminds me why i don't willingly use c++ ;) 2011-09-03T10:16:31 it's such a long-winded way of saying bijection : Game u p -> Game u p -> Bijection u p 2011-09-03T10:36:07 *** dvladim has joined #aichallenge 2011-09-03T10:44:21 *** ibdknox has joined #aichallenge 2011-09-03T10:52:10 *** delt0r_ has quit IRC (Ping timeout: 260 seconds) 2011-09-03T11:03:48 *** sigh has quit IRC (Remote host closed the connection) 2011-09-03T11:05:20 *** delt0r_ has joined #aichallenge 2011-09-03T11:31:16 *** deeper-blue has joined #aichallenge 2011-09-03T12:11:41 *** kaemo has quit IRC (Ping timeout: 240 seconds) 2011-09-03T12:13:52 not sure i like the replay format 2011-09-03T12:14:10 but hey since the nice shinny view works so well with it 2011-09-03T12:45:59 *** dvladim has quit IRC (Ping timeout: 245 seconds) 2011-09-03T12:50:15 i think the stream format still works with the visualizer, and it's more close to what you would capture from the bot i/o 2011-09-03T13:11:44 *** eashoka has quit IRC (Quit: Page closed) 2011-09-03T13:33:42 berak: Stream format? 2011-09-03T13:36:43 playgame -S trying to pastebin an example 2011-09-03T13:38:44 well want to keep my stuff as compatible as possible of course 2011-09-03T13:43:07 *** kaemo has joined #aichallenge 2011-09-03T13:47:29 berak: what format does the visualizer use? both? 2011-09-03T13:50:23 i think it's trying to support anything it it encountered in it's long run .. 2011-09-03T13:54:04 but right, that pulls additional burdon on poor frontier.. 2011-09-03T13:58:04 ok, and it didn't work, so sorry for that 2011-09-03T14:11:09 *** rabidus has quit IRC (Ping timeout: 240 seconds) 2011-09-03T14:12:19 *** rabidus has joined #aichallenge 2011-09-03T14:16:10 *** mathis_ has quit IRC (Ping timeout: 252 seconds) 2011-09-03T14:18:50 *** mathis_ has joined #aichallenge 2011-09-03T14:27:32 *** mathis_ has quit IRC (Ping timeout: 252 seconds) 2011-09-03T14:29:09 *** mathis_ has joined #aichallenge 2011-09-03T14:35:06 *** Cyndre_ has quit IRC (Ping timeout: 258 seconds) 2011-09-03T14:39:25 jmcarthur: :P 2011-09-03T14:40:26 I shouldn't show you this (http://pastebin.com/Jvsuy3WC) then :P 2011-09-03T14:41:01 that's for iterating through elements of a subgroup of Sn :\ 2011-09-03T14:55:44 omg 2011-09-03T14:57:11 it at least works :P 2011-09-03T14:57:33 i really wish /circ was on keyboards :( 2011-09-03T14:57:56 i don't like using * for both . and /circ 2011-09-03T14:58:13 what editor do you use? 2011-09-03T14:58:18 code::blocks 2011-09-03T14:58:31 ah 2011-09-03T14:58:34 i should really learn vim but meh :P 2011-09-03T14:58:44 if you said emacs i would have recommended you check out how agda-mode does unicode input 2011-09-03T14:58:52 or that :P 2011-09-03T14:59:07 in agda-mode, i can just type \circ :) 2011-09-03T14:59:24 or \comp 2011-09-03T14:59:35 hmmm, that is quite appealing 2011-09-03T15:00:41 at one time i had pulled that input mode out to other emacs modes, but i don't remember how anymore 2011-09-03T15:02:34 this is all fairly pointless anyway, the number of bijections between a game explodes as the number of players and strategies increase, but I have a better way for finding them anyway I think, just want to implement this stuff anyway 2011-09-03T15:03:32 *** dvladim has joined #aichallenge 2011-09-03T15:04:20 code like that is begging for list comprehensions ;) 2011-09-03T15:05:33 i admit i don't really understand what the algorithm does 2011-09-03T15:05:52 do you know the algorithm to lexicographically iterate through S_n? 2011-09-03T15:05:58 http://en.wikipedia.org/wiki/Permutation#Systematic_generation_of_all_permutations 2011-09-03T15:06:05 it's a generalisation of that 2011-09-03T15:06:35 oh 2011-09-03T15:06:41 where I only want a subgroup of S_n which maps between players of two games that have the same number of strategies 2011-09-03T15:06:46 it's even easier than i thought in haskell then 2011-09-03T15:07:04 just use the permutations function to get a lazy list of permutations :D 2011-09-03T15:07:30 I imagine the permutations function is that? 2011-09-03T15:07:51 i'm not sure what the algorithm is, but i know it's lazy (doesn't generate all permutations at once) 2011-09-03T15:08:06 and i know it doesn't duplicate work unnecessarily 2011-09-03T15:08:16 well 2011-09-03T15:08:20 i don't know that, actually 2011-09-03T15:08:24 it might not be that efficient algorithm 2011-09-03T15:08:49 ah, indeed it seems not to be 2011-09-03T15:08:50 yeah this will iterate from the identity through the elements lexicographically (and I wrap around at the end back to the identity) so you can generate it without repeats 2011-09-03T15:08:55 hmm okay :\ 2011-09-03T15:09:05 well, it doesn't repeat 2011-09-03T15:09:15 *** ibdknox has joined #aichallenge 2011-09-03T15:09:22 but i just tested in ghci and it slows down as you go, when there are tons of permutations at least 2011-09-03T15:10:03 my bijections are pretty much a semidirect product, but I can't quite get it defined as such when I treat plays strategy sets as sets rather than just Z_di 2011-09-03T15:10:04 *** Cyndre has joined #aichallenge 2011-09-03T15:10:23 bear in mind there's n! permutations :) 2011-09-03T15:10:48 if I have d strategies for each player there's (n!)(n)(d!) game bijections :| 2011-09-03T15:10:48 i realize that, but i mean it gets slower and slower at generating each next permutation for the same inut set 2011-09-03T15:10:51 *input 2011-09-03T15:10:55 ah yeah okay 2011-09-03T15:11:32 hmm 2011-09-03T15:11:40 maybe i jumped the gun there 2011-09-03T15:11:44 i'll just look at the source 2011-09-03T15:12:36 tratie men keyboard kapotgemept 2011-09-03T15:12:40 i have an alternative idea for testing ordinal equivalence where the "number of cases" increases the further players ordinal preferences over the strategy space is from a partial order (it's a weak order) 2011-09-03T15:13:01 ignore that 2011-09-03T15:13:40 which is basically take an ordered copy of the strategy space for each player then start matching up strategies until you get a contradiction 2011-09-03T15:14:29 you can do something similar for cardinal equivalence and keeping a min/max for a_i and b_i for the strictly monotonic linear transformation 2011-09-03T15:14:34 i don't understand this permutations function :( 2011-09-03T15:14:47 link? 2011-09-03T15:15:01 http://www.haskell.org/ghc/docs/latest/html/libraries/base-4.4.0.0/src/Data-List.html#permutations 2011-09-03T15:15:17 it does look like an algorithm that is insufficiently incremental 2011-09-03T15:15:37 ... and that's why I don't use haskell (yet) as I don't understand it :P 2011-09-03T15:15:42 it can lazily generate the permutations, but it seems to get slower as you go 2011-09-03T15:15:58 oh it's not the code i don't understand. it's just the algorithm 2011-09-03T15:16:11 yeah, I was stating that for me :) 2011-09-03T15:16:33 ummm.. can you in some way tell a list to generate iteratively lazily and it'll assume you are going to access it iteratively? 2011-09-03T15:16:44 cause then you could do this other algorithm for iterating through them 2011-09-03T15:17:43 yes 2011-09-03T15:17:47 that's basically the default 2011-09-03T15:18:10 the algorithm you have seems to rely on mutation, though i suspect there may be a purely functional variant 2011-09-03T15:18:20 or you could just use mutation behind the scenes 2011-09-03T15:18:30 what's mutation? 2011-09-03T15:18:38 destructive assignment 2011-09-03T15:19:03 overwriting one value with another 2011-09-03T15:19:12 ah yeah okay 2011-09-03T15:20:22 *** zumbert has joined #aichallenge 2011-09-03T15:20:33 I think i'm basically saying my subgroup of Sn is isomorphic to (S_d1x...S_dn)x(...) where each d1 = ... = dn for each () part :\ 2011-09-03T15:20:58 then using a partition of the player sets to cycle through each (Sd1x...xSdn 2011-09-03T15:21:32 that sounds very large 2011-09-03T15:22:04 it's HUGE, but I have alternative ideas for constructing a bijection based on orderings of the strategy sets from players preferences :) 2011-09-03T15:22:54 i had this elaborate plan for my bot, but i've lost interest since things are moving so slowly 2011-09-03T15:23:01 *** dvladim has quit IRC (Ping timeout: 258 seconds) 2011-09-03T15:23:08 but i at least need a way for testing ordinal equivalence of 2x2 games (where there's only 8 possible bijections) so I can count the number of equivalence classes for structurally different 2x2 games based on ordinal equivalence of players utility functions :\ 2011-09-03T15:23:19 this isn't for the ai contest, this is for my honours thesis :) 2011-09-03T15:23:25 ha 2011-09-03T15:23:29 fine then 2011-09-03T15:24:33 hopefully the contest gets finished, but i have neither the time nor the ability to fix what needs fixing :\ 2011-09-03T15:24:53 although at this rate i'm beginning to hope it doesn't get fixed until i finish uni for the year :P 2011-09-03T15:24:58 so i can actually compete 2011-09-03T15:27:28 antimatroid: are you aware that the C++ STL has a next_permutation function? 2011-09-03T15:27:51 ehe, nope :P 2011-09-03T15:28:00 http://www.cplusplus.com/reference/algorithm/next_permutation/ 2011-09-03T15:28:00 i have fun reinventing the wheel though 2011-09-03T15:28:19 it uses the knuth algorithm 2011-09-03T15:28:34 err, not sure if knuth actually invented that one or not, actually 2011-09-03T15:28:58 wouldn't surprise me 2011-09-03T15:29:05 actually, he didn't 2011-09-03T15:29:16 i think it says it's been "reinvented" a number of times independently 2011-09-03T15:29:20 starting ages ago 2011-09-03T15:29:24 yeah something like that 2011-09-03T15:29:40 " The method goes back to Narayana Pandita in 14th century India, and has been frequently rediscovered ever since." 2011-09-03T15:29:43 that's the comment 2011-09-03T15:30:06 indians did a lot 2011-09-03T15:30:35 http://en.wikipedia.org/wiki/Srinivasa_Ramanujan cough 2011-09-03T15:30:56 and basically started algebra 2011-09-03T15:32:16 "When asked about the methods employed by Ramanujan to arrive at his solutions, Hardy said that they were "arrived at by a process of mingled argument, intuition, and induction, of which he was entirely unable to give any coherent account."" ehehe 2011-09-03T15:34:26 *** ibdknox has quit IRC (Remote host closed the connection) 2011-09-03T15:35:06 *** deeper-blue has quit IRC (Quit: deeper-blue) 2011-09-03T15:41:44 *** berak has quit IRC (Quit: ChatZilla 0.9.84 [SeaMonkey 2.0a3/20090223135443]) 2011-09-03T15:43:04 *** onensora has joined #aichallenge 2011-09-03T15:43:55 *** Eruonen has quit IRC (Ping timeout: 250 seconds) 2011-09-03T15:53:32 *** Accoun has quit IRC () 2011-09-03T16:05:39 *** Accoun has joined #aichallenge 2011-09-03T16:14:40 *** deeper-blue has joined #aichallenge 2011-09-03T16:15:20 *** zumbert has quit IRC (Ping timeout: 252 seconds) 2011-09-03T16:45:46 *** mcstar has joined #aichallenge 2011-09-03T16:51:43 antimatroid: i have an example for creating all the subsets of a set 2011-09-03T16:51:55 http://pastebin.com/dEDAr5rw 2011-09-03T16:52:03 it uses binary counting 2011-09-03T16:52:27 the power set :P 2011-09-03T16:52:31 yep 2011-09-03T16:53:25 antimatroid: my reason for staying away from haskell is the same as yours :D 2011-09-03T16:53:47 its just seems very complicated(the code) 2011-09-03T16:53:54 -s 2011-09-03T16:59:28 *** heinrich5991 has quit IRC (Ping timeout: 264 seconds) 2011-09-03T17:05:06 *** heinrich5991 has joined #aichallenge 2011-09-03T17:22:07 *** kaemo0 has joined #aichallenge 2011-09-03T17:22:50 *** kaemo has quit IRC (Ping timeout: 268 seconds) 2011-09-03T17:29:21 *** ibdknox has joined #aichallenge 2011-09-03T17:31:32 *** mceier has quit IRC (Quit: leaving) 2011-09-03T17:45:02 *** amstan has joined #aichallenge 2011-09-03T17:45:02 *** ChanServ sets mode: +o amstan 2011-09-03T17:55:56 *** mathis_ has quit IRC (Quit: Page closed) 2011-09-03T18:01:19 gn 2011-09-03T18:01:21 *** mcstar has quit IRC (Quit: WeeChat 0.3.5) 2011-09-03T18:03:29 *** amstan has quit IRC (Ping timeout: 245 seconds) 2011-09-03T18:17:30 dang, and i wanted to show off my haskell powerset function to mcstar 2011-09-03T18:17:52 filterM (const [True, False]) 2011-09-03T18:19:04 *** deeper-blue has quit IRC (Quit: deeper-blue) 2011-09-03T18:19:29 *** m__ has joined #aichallenge 2011-09-03T18:19:35 perhaps not as efficient as his though :) 2011-09-03T18:25:41 *** m__ has quit IRC (Quit: Page closed) 2011-09-03T18:51:59 *** delt0r_ has quit IRC (Ping timeout: 252 seconds) 2011-09-03T19:03:10 *** deeper-blue has joined #aichallenge 2011-09-03T19:05:37 *** delt0r_ has joined #aichallenge 2011-09-03T19:23:03 *** FireFly has quit IRC (Quit: FireFly) 2011-09-03T19:55:51 *** Kingpin13 has quit IRC (Quit: quit) 2011-09-03T20:31:09 *** kaemo0 has quit IRC (Ping timeout: 240 seconds) 2011-09-03T20:38:02 *** amstan has joined #aichallenge 2011-09-03T20:38:02 *** ChanServ sets mode: +o amstan 2011-09-03T20:41:08 *** Lithosphere has quit IRC (Remote host closed the connection) 2011-09-03T20:54:27 delt0r_: ping 2011-09-03T20:57:15 *** mleise has quit IRC (Quit: Leaving.) 2011-09-03T20:57:36 *** mleise has joined #aichallenge 2011-09-03T21:01:49 delt0r: berak is right, the I try to support all formats that were ever used by the official engine. That's not a problem though, because by just supporting 'storage', 'streaming' and 'map' formats I can reuse that old code well. 2011-09-03T21:02:36 *** Redgis has quit IRC (Ping timeout: 255 seconds) 2011-09-03T21:03:18 The 'storage' replay format was our attempt to a) get a common data exchange format into the game (JSON) and b) reduce the file size. 2011-09-03T21:05:12 If you look at the file sizes of the 'streaming' and 'storage' format you see why. Also I've added a play_one_game_live.sh/cmd to the tools to make it easier to use the engine's streaming output with the visualizer. 2011-09-03T21:20:25 mleise: well the storage format is nice in that you have info about what each ant does 2011-09-03T21:20:52 while you only have the state at each turn with streaming 2011-09-03T21:21:42 well i will work on it soon. 3am here --summer parties are great ;) 2011-09-03T21:34:09 can't wait for summer :) 2011-09-03T21:44:31 *** ibdknox has quit IRC (Remote host closed the connection) 2011-09-03T21:59:45 *** sigh has joined #aichallenge 2011-09-03T22:09:24 *** hearGrm has joined #aichallenge 2011-09-03T22:33:10 D: summer is ending in canada 2011-09-03T22:42:32 *** hearGrm has quit IRC (Ping timeout: 252 seconds)