2011-09-09T00:06:12 *** McLeopold has left #aichallenge 2011-09-09T00:08:37 *** bhasker has quit IRC (Quit: bhasker) 2011-09-09T01:14:26 *** amstan has quit IRC (Ping timeout: 260 seconds) 2011-09-09T01:22:35 *** ArchMonkey has quit IRC (Quit: ArchMonkey) 2011-09-09T02:07:37 *** jako has joined #aichallenge 2011-09-09T02:37:41 *** delt0r_ has quit IRC (Ping timeout: 240 seconds) 2011-09-09T02:39:25 *** stocha has joined #aichallenge 2011-09-09T02:39:29 delt0r: you'll have to talk to McLeopold about food spawn rate, he worked it out a while ago 2011-09-09T02:39:42 now, even if the ant challenge do launch, there will be a problem 2011-09-09T02:39:57 early beta participants had months to debug their stuff. 2011-09-09T02:40:16 if the duration is 3 month, that will be quite unfair for those starting at launch time. 2011-09-09T02:40:25 if it is less, it will be even worst 2011-09-09T02:41:19 stocha: there are always disadvantages, this is perhaps undesirable but nothing can be done about it 2011-09-09T02:41:20 so 4 months or 5 month would be more reasonable 2011-09-09T02:41:26 that's too long 2011-09-09T02:41:37 i do think we should go for 3 rather than 2 given how long beta has been though 2011-09-09T02:41:47 when beta tester had a year to build up their thing, 5 month really is more fair. 2011-09-09T02:42:22 okay 3 still give a bit of room for maneuvre. But i think it has to be more now. 2011-09-09T02:43:19 i didn't get involved much, so the people who have been improving their bot more regularly have more data to judge. 2011-09-09T02:43:59 but the thing is, with more time, even you don't work on it full time, you'll have more ideas. Not mentioning that you do have more time also. 2011-09-09T02:44:23 anyway it's still not clear if the thing will launch eventually. 2011-09-09T02:50:09 *** delt0r_ has joined #aichallenge 2011-09-09T02:50:36 delt0r_: i'm trying to work out this spawn formula again, but it's confusing me :P 2011-09-09T02:51:10 its not clear from the code... well not after about 10min... i have not looked again 2011-09-09T02:51:43 I was going to just do symmetric... bassed on start positions 2011-09-09T02:51:46 i think i can explain it, but it'd be worth checking with mcleopold later (i can't explain the implementation, only what it's meant to be doing) 2011-09-09T02:53:09 delt0r_: do you understand what I mean when I say take a generating set of locations for the map under symmetry? 2011-09-09T02:53:54 it's a set of locations that spans the set of locations when you take all symmetrically equivalent locations and no two locations in the generating set are symmetrically equivalent 2011-09-09T02:53:56 not really... if its always folded symmetry yes... but what about point symmetry? 2011-09-09T02:54:05 rotational I mean 2011-09-09T02:54:16 aka like we had in planet wars 2011-09-09T02:54:25 same thing, each square has <= n symmetrically equivalent locations including itself right? 2011-09-09T02:54:46 yea 2011-09-09T02:54:56 you mean generalized symmetry 2011-09-09T02:55:43 yeah might as well go with that, it holds for any, you find a set of squares that aren't symmetrically equivalent to each other but if you include all symmetrically equivalent locations then you get the entire set of locations 2011-09-09T02:56:54 then you randomly order your generating set and spawn food in that order for the locations, if you spawn food at a location in the generating set then you also spawn food at every location symmetrically equivalent to it 2011-09-09T02:57:37 if a square picked to spawn food on is not free to have food, the engine waits to spawn food there the next turn it's available to 2011-09-09T02:58:13 that last bit is the best we've come up with to try and make sure the actions of a bot can't influence food spawning on another part of the map 2011-09-09T02:58:22 only hurt themselves by not moving and getting the food earlier 2011-09-09T02:58:51 are you with me so far? the last bit to explain is how the spawn rate works, but i'd like to work the formula out again before explaining that :) 2011-09-09T03:03:27 delt0r_: using the get_symmetric_loc function it's easy to get the generating set, pick a location add it to the generating set and mark all symmetrically equivalent locations as visited, then bfs outwards finding locations not found yet doing a similar thing 2011-09-09T03:04:12 *** Palmik has joined #aichallenge 2011-09-09T03:05:01 Yea i follow 2011-09-09T03:05:03 *** stocha has quit IRC (Quit: Page closed) 2011-09-09T03:05:26 what is the food production rate? 2011-09-09T03:05:37 numberPlayers*food_rate? 2011-09-09T03:05:53 and is food rate fractional... aka .5 per turn 2011-09-09T03:06:22 is this interpreted as a probability if it is fractional 2011-09-09T03:07:00 Finally with scores they seem to be integers. Yet you can have fractional scores --why not have score as a real? 2011-09-09T03:09:03 nah it's a bit more complicated i think 2011-09-09T03:09:48 let f be the desired food spawn rate per turn, but that's f food squares for all players and not per player, n the number of players and t the turn, I think foodToSpawn = (f + (t-1)*f%n)%n but i need to check that 2011-09-09T03:09:49 hang on 2011-09-09T03:10:12 why...why make it so complicated... this is only going to have bugs and require changing to the cries of "i would have won if the rules didn't change" 2011-09-09T03:10:51 a simple probabilistic model would have the same average and be simpler... 2011-09-09T03:11:20 some locations don't have n symmetrically equivalent locations including itself depending on the symmetry 2011-09-09T03:11:36 and now you can have a food spawn rate not as a multiple of the number of players 2011-09-09T03:11:52 oh, that's not right either :P 2011-09-09T03:12:21 so 2011-09-09T03:12:36 the spawn rate is a little different map to map... 2011-09-09T03:13:08 also that function has the symmetry specified... ie rotational or tile 2011-09-09T03:13:34 how? 2011-09-09T03:13:44 oh get_symmetric_locs? 2011-09-09T03:13:57 yea 2011-09-09T03:14:00 I c 2011-09-09T03:14:02 yeah we need to change over to carlop's general symmetry but haven't 2011-09-09T03:14:09 get_symmetric_loc 2011-09-09T03:14:13 is above it 2011-09-09T03:14:14 but it's just a matter of calling the right function for the given symmetry 2011-09-09T03:14:35 just assume tile and folded symmetry 2011-09-09T03:14:53 assumes 2011-09-09T03:16:28 i think it's foodToSpawn = (f + (t-1)*f%n) - (f + (t-1)*f%n)%n 2011-09-09T03:16:35 that's gross, i swear there's a simpler way to do it 2011-09-09T03:17:18 i think everyone is trying to hard for "perfect" 2011-09-09T03:17:39 just have a stochastic rate target... 2011-09-09T03:17:56 what is f t and n ? 2011-09-09T03:18:12 t is turn, n is number of players and f is the spawn rate 2011-09-09T03:19:04 so say you have n = 5 and f = 7 2011-09-09T03:19:05 nf = {0, 7, 14, 21, 28, 35, 42, ...} 2011-09-09T03:19:05 toSpawn = {0, 5, 5, 10, 5, 10, 5, ...} 2011-09-09T03:19:20 that doesn't allow for squares with < n symmetrically equivalent locations aye 2011-09-09T03:19:21 hmmmm 2011-09-09T03:19:54 errr, that's tf sorry 2011-09-09T03:20:48 *** ltriant has quit IRC (Quit: Get MacIrssi - http://www.sysctl.co.uk/projects/macirssi/) 2011-09-09T03:21:21 delt0r_: personally I would like the symmetry to go so far as that a deterministic bot playing itself will always draw 2011-09-09T03:21:53 *** aerique has joined #aichallenge 2011-09-09T03:22:11 although it's harder to make a bot play like that than one would probably first think because of the wrapping 2011-09-09T03:22:42 I work on stochastic models.. I only need the expectation to be as many wins as losses 2011-09-09T03:23:03 in the limit of infinite games played ;) 2011-09-09T03:23:39 personally i want to minimise the amount of "noise" for ranking bots because there's already so much 2011-09-09T03:24:00 and i've probably had quite a bit of influence on how stuff is done, if it can be done reasonably well easier i'm happy to change 2011-09-09T03:24:14 we'd need to get mcleopold to agree as well though 2011-09-09T03:24:29 i do like the ordering the generating set for finding locations to spawn on though 2011-09-09T03:24:54 yea-- but what does that formular even mean... why not just x number of food per turn 2011-09-09T03:25:38 because say you have f = 7 and n=5 with all squares having n symmetrically equivalent locations, you can't spawn 7 food locations fairly each turn 2011-09-09T03:26:21 you going to get a lot of question on the why --lots 2011-09-09T03:26:41 oh ffs 2011-09-09T03:27:20 yeah, but i can write up clearer explanations of things like that once stuff is finalised and we're ready to launch 2011-09-09T03:27:29 then just point people to that, the same is going to happen for battle resolution 2011-09-09T03:27:31 use the closest value that is a multiple of the number of players... 2011-09-09T03:27:47 and i think fewer people will care about the food spawning method other than being generally told it's "fair" 2011-09-09T03:27:55 no they won't 2011-09-09T03:27:58 yeah, that would work too 2011-09-09T03:28:10 they won't believe you unless they understand it 2011-09-09T03:28:24 see he who shall not be named 2011-09-09T03:48:39 *** onensora has quit IRC (Ping timeout: 252 seconds) 2011-09-09T03:55:11 *** mleise has quit IRC (Ping timeout: 240 seconds) 2011-09-09T04:12:57 *** boegel has joined #aichallenge 2011-09-09T04:14:26 *** berak has joined #aichallenge 2011-09-09T04:39:26 *** TTE has joined #aichallenge 2011-09-09T04:53:03 *** sigh has joined #aichallenge 2011-09-09T05:06:23 *** mceier has joined #aichallenge 2011-09-09T05:11:53 *** antimatroid has quit IRC (Ping timeout: 252 seconds) 2011-09-09T05:20:35 *** antimatroid has joined #aichallenge 2011-09-09T06:13:21 hey... cool map... the combat map--each get 10 ants... no food... battle to the death. :D 2011-09-09T06:13:42 for the lutz 2011-09-09T06:26:37 do the example bots try to move onto water? 2011-09-09T06:32:53 And what way is north? col++ or row++ 2011-09-09T06:32:56 or -- 2011-09-09T06:38:24 delt0r: row++ is north? 2011-09-09T06:38:28 how could it be anything but? 2011-09-09T06:38:41 delt0r_* 2011-09-09T06:38:59 well I use x and y 2011-09-09T06:39:10 rows are for wimps and communists 2011-09-09T06:39:11 ;) 2011-09-09T06:39:15 hmmm, i still maintain that's inferior :P 2011-09-09T06:39:27 it's a pita when formatting input or output 2011-09-09T06:39:27 x and y and right dam handed 2011-09-09T06:39:35 x goes up and down 2011-09-09T06:39:46 ehe :) 2011-09-09T06:39:51 no--rows are a pita 2011-09-09T06:39:54 and cols 2011-09-09T06:39:57 why? 2011-09-09T06:39:58 matrices! 2011-09-09T06:40:05 are we row major or row minor? 2011-09-09T06:40:34 everything has a different convention --with a right handed coordinate system is very convention 2011-09-09T06:40:45 anyway 2011-09-09T06:40:48 most people treat (x,y) like when drawing graphs 2011-09-09T06:40:52 I have a bug i think... 2011-09-09T06:40:56 yea 2011-09-09T06:41:05 convention that is kept 2011-09-09T06:41:31 while R matlab opengl etc all have different row major or col major matrix formats 2011-09-09T06:42:06 meh :P 2011-09-09T06:42:29 I don't think the random bot would move onto water... but it does in my engine... hence i think i have a bug 2011-09-09T06:42:48 i'm not sure whether it does or doesn't 2011-09-09T06:42:54 i can tell you if my starter bot does 2011-09-09T06:43:02 from the code it looks like no 2011-09-09T06:43:13 could be that i give it bogus water data 2011-09-09T06:43:27 * delt0r_ gets lunch 2011-09-09T06:44:34 yeah the c++ starter bot avoids water 2011-09-09T07:03:16 *** antimatroid has quit IRC (Ping timeout: 260 seconds) 2011-09-09T07:23:08 I can't get any java bot to run with the playgame.py script 2011-09-09T07:23:19 the commant works fine at the shell 2011-09-09T07:23:36 but as a bot argument i get nothing 2011-09-09T07:23:50 the error message with -E is "the bot crashed" 2011-09-09T07:24:10 there is no -O for the bot... 2011-09-09T07:24:22 is python mangling the command line or something? 2011-09-09T07:25:37 what does your commandline look like? 2011-09-09T07:27:38 oh wait, java. nevermind. i've only used the python sample bots 2011-09-09T07:28:05 ok --just got some action with a bash shell script 2011-09-09T07:28:21 but directly calling java does not seem to work 2011-09-09T07:48:54 found my engine bug 2011-09-09T07:49:03 was not waiting for go before the first turn 2011-09-09T07:49:07 from the bots 2011-09-09T07:55:38 The GreedyBot.py times out on big maps with lots of ants 2011-09-09T07:55:55 only had 1000 ants 2011-09-09T08:14:55 well we are getting there... so there may be a BorgAntManager this weekend yet 2011-09-09T08:29:42 *** onensora has joined #aichallenge 2011-09-09T08:33:04 *** mathis has joined #aichallenge 2011-09-09T08:39:31 \o/ 2011-09-09T09:06:48 *** mceier has quit IRC (Ping timeout: 260 seconds) 2011-09-09T09:44:05 will the focus battle resolution scheme be used for the contest? has this already been decided? 2011-09-09T10:19:35 jako: judging by how keen some folks are to just start --i would say that its the one that will get picked... 2011-09-09T10:23:02 *** boegel has quit IRC (Quit: *poof!*) 2011-09-09T10:29:51 *** sigh has quit IRC (Read error: Connection reset by peer) 2011-09-09T10:34:50 *** jako has quit IRC (Ping timeout: 252 seconds) 2011-09-09T10:35:42 *** delt0r_ has quit IRC (Read error: Operation timed out) 2011-09-09T10:45:17 *** jako has joined #aichallenge 2011-09-09T10:50:37 *** delt0r_ has joined #aichallenge 2011-09-09T10:54:53 *** aerique has quit IRC (Quit: ...) 2011-09-09T10:56:52 *** mcstar has joined #aichallenge 2011-09-09T10:57:42 fewer people than ever 2011-09-09T11:06:21 *** ArchMonkey has joined #aichallenge 2011-09-09T11:07:16 *** ArchMonkey has quit IRC (Client Quit) 2011-09-09T11:17:01 *** antimatroid has joined #aichallenge 2011-09-09T11:19:11 *** FireFly has joined #aichallenge 2011-09-09T11:28:36 *** dvladim has joined #aichallenge 2011-09-09T11:30:11 *** mcstar has quit IRC (Quit: WeeChat 0.3.5) 2011-09-09T12:00:46 *** nux67 has quit IRC (Quit: Page closed) 2011-09-09T12:18:53 *** dvladim has quit IRC (Ping timeout: 276 seconds) 2011-09-09T12:28:09 *** mceier has joined #aichallenge 2011-09-09T12:39:26 *** Eruonen has joined #aichallenge 2011-09-09T12:42:31 *** onensora has quit IRC (Ping timeout: 252 seconds) 2011-09-09T12:55:08 *** McLeopold has joined #aichallenge 2011-09-09T12:58:13 *** mceier has quit IRC (Ping timeout: 260 seconds) 2011-09-09T13:00:00 *** mceier has joined #aichallenge 2011-09-09T13:00:25 *** rabidus has quit IRC (Read error: Connection reset by peer) 2011-09-09T13:01:03 *** rabidus has joined #aichallenge 2011-09-09T13:18:10 *** dvladim has joined #aichallenge 2011-09-09T13:26:09 *** sir_macelon has joined #aichallenge 2011-09-09T14:00:12 *** mceier has quit IRC (Read error: Operation timed out) 2011-09-09T14:02:04 *** amstan has joined #aichallenge 2011-09-09T14:02:04 *** ChanServ sets mode: +o amstan 2011-09-09T14:04:20 *** mceier has joined #aichallenge 2011-09-09T14:15:58 *** dvladim has quit IRC (Ping timeout: 260 seconds) 2011-09-09T14:19:14 *** mcstar has joined #aichallenge 2011-09-09T14:26:58 *** mathis has quit IRC (Quit: Page closed) 2011-09-09T14:28:03 *** Cyndre has quit IRC (Quit: Leaving) 2011-09-09T14:28:14 *** mathis has joined #aichallenge 2011-09-09T14:37:53 *** mceier has quit IRC (Read error: Operation timed out) 2011-09-09T14:38:34 *** rabidus has quit IRC (Ping timeout: 260 seconds) 2011-09-09T14:40:11 janzert: you around? 2011-09-09T14:40:39 amstan: you around? 2011-09-09T14:40:45 ya? 2011-09-09T14:41:09 I'm going to make some major changes to the data, I wanted somebody on standby to restart workers if there are issues 2011-09-09T14:41:24 ok 2011-09-09T14:41:27 since I don't have access... 2011-09-09T14:42:15 *** mceier has joined #aichallenge 2011-09-09T14:44:10 McLeopold: ready 2011-09-09T14:44:43 okay, it might take me 30 minutes or so to do the change 2011-09-09T14:44:47 ok 2011-09-09T14:50:42 *** rabidus has joined #aichallenge 2011-09-09T14:52:04 amstan: I forgot how to get to the worker screen session :( 2011-09-09T14:52:24 there's no worker on the main server 2011-09-09T14:52:36 got it, I wanted to stop the leaderboard updates 2011-09-09T14:53:06 McLeopold: btw, there was an uncommited file 2011-09-09T14:53:11 i had to reset it in order to pull 2011-09-09T14:53:13 yea, that was me 2011-09-09T14:53:21 it was the missing comma commit 2011-09-09T14:53:22 frontier freaked out when i did it 2011-09-09T14:53:52 I just changed it on the server to see if it would work, the made the change locally, pushed, and forgot to clean up the server 2011-09-09T14:56:15 amstan: okay, I'm going to delay my changes another hour 2011-09-09T14:56:21 ok 2011-09-09T14:56:28 my terminal will probably stay up for a while 2011-09-09T14:57:01 the leaderboard will stay turned off for now 2011-09-09T15:16:25 *** AlliedEnvy_ has joined #aichallenge 2011-09-09T15:18:29 *** AlliedEnvy has quit IRC (Ping timeout: 245 seconds) 2011-09-09T15:20:11 *** rabidus has quit IRC (Ping timeout: 240 seconds) 2011-09-09T15:22:15 *** rabidus has joined #aichallenge 2011-09-09T15:22:53 *** AlliedEnvy_ has quit IRC (Ping timeout: 260 seconds) 2011-09-09T15:30:48 *** Accoun has quit IRC () 2011-09-09T15:34:57 *** AlliedEnvy_ has joined #aichallenge 2011-09-09T15:35:03 *** sir_macelon has quit IRC (Quit: Page closed) 2011-09-09T15:44:12 *** mathis has quit IRC (Disconnected by services) 2011-09-09T15:46:48 *** Accoun has joined #aichallenge 2011-09-09T15:51:43 *** AlliedEnvy_ has quit IRC (Ping timeout: 252 seconds) 2011-09-09T16:04:35 *** AlliedEnvy_ has joined #aichallenge 2011-09-09T16:20:44 *** mceier has quit IRC (Ping timeout: 260 seconds) 2011-09-09T16:22:11 *** mceier has joined #aichallenge 2011-09-09T16:28:28 *** berak has quit IRC (Quit: ChatZilla 0.9.84 [SeaMonkey 2.0a3/20090223135443]) 2011-09-09T16:59:19 *** Palmik has quit IRC (Read error: Connection reset by peer) 2011-09-09T17:03:44 *** mcstar has quit IRC (Quit: WeeChat 0.3.5) 2011-09-09T17:03:53 *** Daniela06 has joined #aichallenge 2011-09-09T17:36:23 *** McLeopold1 has joined #aichallenge 2011-09-09T17:38:30 *** McLeopold has quit IRC (Ping timeout: 260 seconds) 2011-09-09T17:40:56 *** McLeopold1 is now known as McLeopold 2011-09-09T17:58:07 aichallenge: McLeopold epsilon * r4b72ff9 / (19 files in 3 dirs): (log message trimmed) 2011-09-09T17:58:07 aichallenge: remove leaderboard table 2011-09-09T17:58:07 aichallenge: Previously rankings were generated once every 60 seconds and the results 2011-09-09T17:58:07 aichallenge: stored in a table. Now the most recent ranking will be stored in the 2011-09-09T17:58:07 aichallenge: submission table and the history will be stored in the game_player 2011-09-09T17:58:08 aichallenge: table. Rank movements for a bot due to other bots changing rank will 2011-09-09T17:58:09 aichallenge: not be captured. The change in ranks between games due to the results 2011-09-09T18:00:43 *** jako has quit IRC (Ping timeout: 252 seconds) 2011-09-09T18:22:54 *** foRei has joined #aichallenge 2011-09-09T18:38:18 *** delt0r_ has quit IRC (Ping timeout: 260 seconds) 2011-09-09T18:50:34 *** delt0r_ has joined #aichallenge 2011-09-09T19:10:43 *** FireFly has quit IRC (Quit: FireFly) 2011-09-09T19:15:57 *** mleise has joined #aichallenge 2011-09-09T19:23:03 aichallenge: McLeopold epsilon * rd4dc481 / (3 files in 2 dirs): 2011-09-09T19:23:03 aichallenge: tweeks to leaderboard table removal 2011-09-09T19:23:03 aichallenge: fix missing comma in sql.php 2011-09-09T19:23:03 aichallenge: fix improper usage of stored proc param 2011-09-09T19:23:03 aichallenge: fix field name in profile.php - http://git.io/cUwbmg 2011-09-09T19:35:43 *** mceier has quit IRC (Quit: leaving) 2011-09-09T19:41:26 aichallenge: McLeopold epsilon * rbcabafd / website/profile.php : unicode/bio fix issue #250 - http://git.io/y3ZdPg 2011-09-09T19:50:27 aichallenge: McLeopold epsilon * r66a21a3 / (website/profile.php website/sql.php): fix sql field name mismatches with profile.php - http://git.io/v2UiJA 2011-09-09T20:01:59 *** McLeopold1 has joined #aichallenge 2011-09-09T20:04:04 *** McLeopold has quit IRC (Ping timeout: 260 seconds) 2011-09-09T20:06:28 aichallenge: McLeopold epsilon * r2f2d702 / (9 files): more unicode utf-8 support - http://git.io/8k2Qlw 2011-09-09T20:09:15 *** McLeopold1 is now known as McLeopold 2011-09-09T20:09:33 aichallenge: McLeopold epsilon * r519f817 / website/sql.php : extra char in sql - http://git.io/ceeGTQ 2011-09-09T20:10:49 *** sigh has joined #aichallenge 2011-09-09T20:18:36 aichallenge: McLeopold epsilon * r764ef74 / website/map.php : fix map.php link, closes #249 - http://git.io/euMs6Q 2011-09-09T20:25:59 amstan, janzert: the leaderboard table is now obsolete, watch for stupid errors in what I did today. 2011-09-09T20:26:05 *** McLeopold has quit IRC (Read error: Connection reset by peer)