2011-06-29T00:02:05 *** davidjliu has quit IRC (Quit: Page closed) 2011-06-29T00:04:41 *** Meatkat has joined #aichallenge 2011-06-29T00:23:05 *** amstan_ has quit IRC (Remote host closed the connection) 2011-06-29T00:23:11 *** amstan_ has joined #aichallenge 2011-06-29T00:23:11 *** ChanServ sets mode: +o amstan_ 2011-06-29T00:23:13 *** amstan_ has quit IRC (Remote host closed the connection) 2011-06-29T00:48:48 *** okayzed is now known as okay 2011-06-29T00:58:58 *** onensora has quit IRC () 2011-06-29T01:15:10 *** McLeopold has quit IRC (Quit: Leaving.) 2011-06-29T01:30:23 *** okay is now known as okayzed 2011-06-29T02:12:02 *** McLeopold has joined #aichallenge 2011-06-29T02:18:21 aichallenge: McLeopold epsilon * r5d55f10 / (41 files in 2 dirs): removed octagonal maps set 1, bad player counts - http://bit.ly/lkLT9V 2011-06-29T02:18:21 aichallenge: McLeopold epsilon * rb483858 / (41 files in 2 dirs): update to map selection - http://bit.ly/jc0f3a 2011-06-29T02:25:15 *** Palmik has joined #aichallenge 2011-06-29T02:28:45 we have a draw! \o/ http://aichallengebeta.hypertriangle.com/visualizer.php?game=44421 2011-06-29T02:31:45 Don't leave your base http://aichallengebeta.hypertriangle.com/visualizer.php?game=44424 2011-06-29T02:33:26 Too much food? http://aichallengebeta.hypertriangle.com/visualizer.php?game=44428 2011-06-29T02:36:27 Fun opening, keep moving http://aichallengebeta.hypertriangle.com/visualizer.php?game=44439 2011-06-29T02:37:44 How to fight and win http://aichallengebeta.hypertriangle.com/visualizer.php?game=44441 2011-06-29T02:38:06 *** McLeopold has left #aichallenge 2011-06-29T02:45:57 *** Meatkat has quit IRC (Read error: Connection reset by peer) 2011-06-29T03:12:00 *** janzert has left #aichallenge 2011-06-29T03:13:59 *** gobang has joined #aichallenge 2011-06-29T03:15:08 *** janzert has joined #aichallenge 2011-06-29T03:19:55 *** aerique has joined #aichallenge 2011-06-29T03:29:44 *** ltriant has quit IRC (Quit: Computer has gone to sleep) 2011-06-29T03:47:28 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-06-29T03:58:10 *** carlop has joined #aichallenge 2011-06-29T04:00:26 *** Accoun has quit IRC (Read error: Connection reset by peer) 2011-06-29T04:03:01 *** noid19987 has joined #aichallenge 2011-06-29T04:03:13 *** noid19987 is now known as Accoun 2011-06-29T04:12:58 @contestbot: later tell McLeopold that I think that map should not have surrounding water (like the octagonal currently up), I think is better having at least 3-4 nearby enemy (instead of 2), cause few near enemy mean that random initial position can heavily influence results (eg 5 strong bot + 3 basic bot, if a strong have near 1/2 basic bot he have huge advantage over a strong bot... 2011-06-29T04:12:59 carlop: I have no idea what you mean. 2011-06-29T04:13:00 ...surrounded by strong bot) 2011-06-29T04:21:55 @contestbot later tell McLeopold that I think that map should not have surrounding water (like the octagonal currently up), I think is better having at least 3-4 nearby enemy (instead of 2), cause few near enemy mean that random initial position can heavily influence results (eg 5 strong bot + 3 basic bot, if a strong have near 1/2 basic bot he have huge advantage over a strong bot surrounded by o 2011-06-29T04:21:56 carlop: User error -- Replace user. 2011-06-29T04:21:57 ther strong bot). For the same reason (have more near team to fight) i think that symmetry like this: http://ants.zeroviz.us/pgEje/ can be better than the octagonal one (I suggest that both should be used) 2011-06-29T04:22:16 contestbot 2011-06-29T04:22:29 help 2011-06-29T04:23:25 @later tell McLeopold that I think that map should not have surrounding water (like the octagonal currently up), I think is better having at least 3-4 nearby enemy (instead of 2), cause few near enemy mean that random initial position can heavily influence results (eg 5 strong bot + 3 basic bot, if a strong have near 1/2 basic bot he have huge advantage over a strong bot surrounded by other... 2011-06-29T04:23:25 carlop: Ready to serve, my lord. 2011-06-29T04:23:27 ...strong bot). For the same reason (have more near team to fight) i think that symmetry like this: http://ants.zeroviz.us/pgEje/ can be better than the octagonal one (I suggest that both should be used) 2011-06-29T04:23:30 finally.... 2011-06-29T04:24:54 *** Naktibalda has joined #aichallenge 2011-06-29T04:34:58 *** boegel has joined #aichallenge 2011-06-29T04:39:19 *** carlop has quit IRC (Remote host closed the connection) 2011-06-29T05:01:56 *** sYnfo has joined #aichallenge 2011-06-29T05:16:19 *** Cyndre_ has quit IRC (Quit: Leaving) 2011-06-29T05:35:57 *** UncleVasya has joined #aichallenge 2011-06-29T05:37:54 Thank you for new maps. Especially for ones that for 2 players. 2011-06-29T05:51:21 *** filter_ is now known as filter 2011-06-29T05:51:55 *** foRei has joined #aichallenge 2011-06-29T05:57:02 UncleVasya: except the new maps suck :P 2011-06-29T05:57:25 why? 2011-06-29T05:59:06 the passages are too small 2011-06-29T05:59:10 and the maps are all bounded 2011-06-29T06:00:56 @later tell McLeopold I'm not sure if this is a bug or me not understanding the engine: https://github.com/aichallenge/aichallenge/issues/204 2011-06-29T06:00:56 aerique: Yep. 2011-06-29T06:01:26 aerique: i'm finding your explanation hard to read :P 2011-06-29T06:01:32 what's the problem? 2011-06-29T06:04:05 antimatroid: i suck at communicating yes :) 2011-06-29T06:04:29 antimatroid: copy & paste the replay and watch it while I try to formulate it again 2011-06-29T06:04:42 aerique: just say what's happening"? 2011-06-29T06:04:54 what moves are you making with the left and right b ant? 2011-06-29T06:04:59 and what's happening? 2011-06-29T06:05:16 giving me rows and columns on a grid i'm looking at doesn't really help :P 2011-06-29T06:05:21 antimatroid: ants 14,11 and 13,11 collide (due to 14,11 moving north and 13,11 having newly spawned) 2011-06-29T06:05:29 which ants, left and right 2011-06-29T06:05:39 locations in a grid aren't helpful when i'm not a computer :P 2011-06-29T06:05:40 antimatroid: right / b 2011-06-29T06:05:52 antimatroid: they move left 2011-06-29T06:05:59 antimatroid: towards the food 2011-06-29T06:06:09 sure.. then? 2011-06-29T06:06:11 antimatroid: new ant spawns, it moves north 2011-06-29T06:06:25 antimatroid: (towards the food above it) 2011-06-29T06:06:40 antimatroid: newly ant has spawned there already so they collide 2011-06-29T06:06:47 antimatroid: this is all as it is supposed to 2011-06-29T06:06:51 wait waht? 2011-06-29T06:06:54 *** sigh has joined #aichallenge 2011-06-29T06:06:59 you move the new ant onto the food location? 2011-06-29T06:07:14 either that ant will die, or it's move will be considered invalid and not processed 2011-06-29T06:07:22 that's what should happen anyway 2011-06-29T06:07:33 antimatroid: yeah, so far so good 2011-06-29T06:07:44 which does it do though? 2011-06-29T06:07:50 does it cancel the move or kill it? 2011-06-29T06:07:55 antimatroid: kill 2011-06-29T06:08:03 so what's the problem? 2011-06-29T06:08:21 so the most left 'b' ant move w,w,w,n and dies colliding with the newly spawned ant 2011-06-29T06:08:38 eh no 2011-06-29T06:08:40 wait, why does it go w, w, w, n? 2011-06-29T06:08:50 you go w, w, then the new ant goes north? 2011-06-29T06:08:53 and new ant dies? 2011-06-29T06:08:56 antimatroid: don't ask why, that's not the point 2011-06-29T06:09:10 reset 2011-06-29T06:09:17 let me make a slideshow... 2011-06-29T06:09:20 the most left b ant moves w, w, w 2011-06-29T06:09:39 or rather w, w at which point a new ant spawns 2011-06-29T06:10:04 that new ant moves n and collides with the new ant that has spawned north of it 2011-06-29T06:10:20 aerique: http://pastebin.com/72gF6g97 2011-06-29T06:10:23 sigh.. yeah do a slide show, this sucks in words 2011-06-29T06:10:27 that's what i'm getting from your explanation? 2011-06-29T06:10:59 erm, does the food die when you collide with it? 2011-06-29T06:11:01 or survive? 2011-06-29T06:11:03 can i add to it? that's not what's happening 2011-06-29T06:11:20 give me something like that :P 2011-06-29T06:11:31 yeah, moment 2011-06-29T06:13:59 http://pastebin.com/Lwqt8Gy2 2011-06-29T06:14:21 changed them to numbers for clarity but they all belong to the same player 2011-06-29T06:14:45 it's blocking your move i suspect 2011-06-29T06:14:55 like if you try to move an ant onto water, it will just ignore your move 2011-06-29T06:15:13 then after it's processed the moves, your ant failed to move, so it's still adjacent to the food, so an ant spawns 2011-06-29T06:15:15 does that make sense? 2011-06-29T06:15:23 Yea, it depends on how you picture the death by merging I guess... 2011-06-29T06:15:31 yeah, that's what i think is happening but should it? 2011-06-29T06:15:38 imo we shouldn't block moves like that, an ant should walk onto water and die 2011-06-29T06:15:48 otherwise it's hard to pick up on a bug where ants try to walk onto water 2011-06-29T06:15:57 but that's what has been done, so meh :) 2011-06-29T06:16:59 i just like the issue to be out there :) 2011-06-29T06:17:09 but thanks for having me clarify it 2011-06-29T06:21:31 *** onensora has joined #aichallenge 2011-06-29T06:34:41 *** mcstar has joined #aichallenge 2011-06-29T06:37:59 hi aerique 2011-06-29T06:38:59 are you interested in a new lambda macro, one where you dont have to explicitely specify the arguments? 2011-06-29T06:39:04 http://pastebin.com/nvVZKpzQ 2011-06-29T06:39:34 im kind of interested, if you could write it in a simpler way... 2011-06-29T06:43:05 *** Zaphus has joined #aichallenge 2011-06-29T06:44:51 *** UncleVasya has quit IRC (Ping timeout: 260 seconds) 2011-06-29T06:52:18 *** Zaphus has quit IRC (Quit: Page closed) 2011-06-29T06:56:30 *** jmreardon has joined #aichallenge 2011-06-29T07:05:12 *** stocha has joined #aichallenge 2011-06-29T07:05:38 *** UncleVasya has joined #aichallenge 2011-06-29T07:06:15 antimatroid: have you seen how much changing the map set affected the ranking ? :) It's probably nothing compared to what will happened if you go for maze maps only though. 2011-06-29T07:07:12 *** sigh has quit IRC (Ping timeout: 252 seconds) 2011-06-29T07:12:11 *** Frontier has joined #aichallenge 2011-06-29T07:13:26 *** Frontier has quit IRC (Client Quit) 2011-06-29T07:15:36 mcstar: have got some examples of its usage? 2011-06-29T07:15:38 @later tell mcleopold: The lesser the radius, the higher the randomness in the games. For example a bot starting with "perfect" food picking on the begining of the game, won't have any advantage if it doesn't know where the food is to begin with. 2011-06-29T07:15:38 stocha: As you wish. 2011-06-29T07:16:29 *** sigh has joined #aichallenge 2011-06-29T07:16:56 hi 2011-06-29T07:17:12 yes, you use it like this: ($ (+ $1 $2)) 2011-06-29T07:17:27 its equivalent to (lambda ($1 $2) (+ $1 $2)) 2011-06-29T07:17:39 and, it can be nested too 2011-06-29T07:18:44 mcstar: i can't see me using it right away but it could be nice to have around. i like args for self-documentation though 2011-06-29T07:19:43 in mathematica you can have a lambda by saying: (#1 + #2)& 2011-06-29T07:19:54 and im quite fond of that 2011-06-29T07:20:37 when you dont want to specify a full-blown lambda-list as an argument to (lambda) i guess its handy that you dont have to write the arguments explicitely 2011-06-29T07:21:06 but its a matter of taste i guess, what im interested in is how would you have written it? 2011-06-29T07:21:17 mcstar: i'm not used to mathematica, but it's great to extend your Lisp like that :) 2011-06-29T07:21:42 mcstar: i very rarely write macros, so better ask on #lisp 2011-06-29T07:21:52 :( 2011-06-29T07:22:04 mcstar: (also i don't have time currently to think about it (am at work)) 2011-06-29T07:22:06 i guess you dont write reader macros either? 2011-06-29T07:22:16 ah, ok 2011-06-29T07:22:21 mcstar: did one last week for the game bot :) http://paste.lisp.org/display/122860 2011-06-29T07:22:21 im not nagging you then 2011-06-29T07:23:40 mcstar: someone on #lisp then pointed me to: file:///export/home/ekwis/emacs/HyperSpec/Body/m_defi_1.htm#define-symbol-macro 2011-06-29T07:25:13 argh, moment 2011-06-29T07:25:39 wheels? deadbody? flies? 2011-06-29T07:25:42 :D 2011-06-29T07:25:49 non-retarded link: http://www.lispworks.com/documentation/HyperSpec/Body/m_defi_1.htm#define-symbol-macro 2011-06-29T07:25:56 ok, i got it already .) 2011-06-29T07:26:15 i use http://xach.com/clhs 2011-06-29T07:26:28 (of course i have a local copy for vim too) 2011-06-29T07:27:34 i see, @ reads the symbol after it, and calls the accessor 2011-06-29T07:28:23 i guess in your case, since you already use globals, you could just have the value, for example turn as a special, dont you? 2011-06-29T07:28:48 *** stocha has quit IRC (Ping timeout: 252 seconds) 2011-06-29T07:30:31 btw i have a simple macro, which sets the function-value of a symbol 2011-06-29T07:30:49 so i use it to bind functions to symbols, so that i dont have to use funcall 2011-06-29T07:31:26 im wondering if setf can be extended, that when it detects a function, it behaves accordingly 2011-06-29T07:31:46 mcstar: something like this?: https://github.com/aerique/okra/blob/master/src-common/common.lisp#L29 2011-06-29T07:32:17 that could perhaps be turned into a defsetf 2011-06-29T07:32:42 like that, but my parameters are exchanged :) 2011-06-29T07:32:57 fail ;-) 2011-06-29T07:33:43 why is it good to have it as a defun? 2011-06-29T07:34:06 i mean theres a difference regarding free variables 2011-06-29T07:34:07 why make it more complicated than necessary? 2011-06-29T07:34:23 you know what i mean? 2011-06-29T07:34:39 no :) 2011-06-29T07:35:43 yeah, no difference 2011-06-29T07:36:28 but with macros. you can eliminate the +1 function call :) 2011-06-29T07:36:41 lame, i know 2011-06-29T07:37:40 i haven't got a good solution for lugging around global state though :-( 2011-06-29T07:39:05 you can pass certain variables as arguments all the time 2011-06-29T07:39:13 i pass rows, cols most of the time 2011-06-29T07:40:30 define-symbol-macro is good for you isnt it? makes those accesses easier 2011-06-29T07:41:00 yup, a little less flexible than the @macro, but probably faster 2011-06-29T07:42:19 if i don't have a global state i get some functions with a great amount of arguments, which is basically the same (and still hard to unit test) except more hidden from view 2011-06-29T07:42:48 guess i should read some books about game programming and see how they solve it 2011-06-29T07:44:00 what i do in my engine, is that i store most of the game variables in a class, and the only important global variable is only that class 2011-06-29T07:44:37 and i initialize it, and after the start-engine returned, i.e. the game terminated for some reason i can still access anything 2011-06-29T07:45:16 i dont think theres a better solution, i guess they use global classess too 2011-06-29T07:45:58 mcstar: same here with the global class 2011-06-29T07:48:27 most of the functions that really depend on that class, are defmethods anyway, so they dont use the global variable, they use their first argument, and other, less connected functions are given the necessary variables as arguments 2011-06-29T07:48:55 only while i prototype something i might use the *board* global, but soon ill modularize it(or functionalize) 2011-06-29T07:50:30 of course this isnt new to you, either :) 2011-06-29T07:50:40 i better start incorporating the new code into my bot 2011-06-29T07:52:46 i see, you are incorporating the utility functions from On Lisp into okra 2011-06-29T07:53:10 yeah :) 2011-06-29T08:02:33 *** dddddddd is now known as davidd 2011-06-29T08:04:50 *** davidd is now known as loolie 2011-06-29T08:04:54 *** loolie is now known as davidd 2011-06-29T08:24:38 *** smiley1983 has joined #aichallenge 2011-06-29T08:45:20 *** sigh_ has joined #aichallenge 2011-06-29T08:46:37 *** sigh__ has joined #aichallenge 2011-06-29T08:47:18 *** sigh has quit IRC (Ping timeout: 252 seconds) 2011-06-29T08:50:14 *** sigh_ has quit IRC (Ping timeout: 250 seconds) 2011-06-29T09:19:03 *** carlop has joined #aichallenge 2011-06-29T09:28:07 *** Kingpin13 has joined #aichallenge 2011-06-29T09:30:44 *** sigh__ has quit IRC (Remote host closed the connection) 2011-06-29T09:32:26 *** thagomizr has quit IRC (Quit: thagomizr) 2011-06-29T09:34:44 *** olexs has joined #aichallenge 2011-06-29T09:39:54 *** stocha has joined #aichallenge 2011-06-29T09:40:06 High there 2011-06-29T09:40:16 j3camero: hi 2011-06-29T09:40:30 j3camero: do you happen to be there ? :) 2011-06-29T09:41:04 j3camero: i'd like to propose 128x128 as a upper bound for maps. So people that need this upper bound can start to build their bots. 2011-06-29T09:41:27 j3camero: i think that higher than 128x128 will get the visualizer into trouble in a lot of cases. 2011-06-29T09:42:04 j3camero: in any way, i'd like an official upper bound for map. 2011-06-29T09:45:29 @later tell McLeopold: hi. I think you were the one to propose 256x256 as a upper bound for maps. What about making 128x128 the official upper bound ? I don't think the visualizer will be able to handle much more in many situations. 2011-06-29T09:45:29 stocha: I come to serve. 2011-06-29T09:45:48 *** stocha has quit IRC (Quit: Page closed) 2011-06-29T09:46:00 stocha: i don;t think upper bound means what you think it does 2011-06-29T09:50:27 *** Accoun has quit IRC (Ping timeout: 240 seconds) 2011-06-29T10:04:32 when tab-completion doesnt work its usually a sign that the person has already left 2011-06-29T10:05:00 (yeah i know, you were typing when he left :)) 2011-06-29T10:05:54 mcstar: you say that like i use tab completion :P 2011-06-29T10:05:59 mcstar: well look at that :P 2011-06-29T10:06:10 my life is going to be a lot easier from now on 2011-06-29T10:06:21 glad that i helped 2011-06-29T10:06:52 i finished my pathfinding 2011-06-29T10:07:02 i mean i finished the prototype 2011-06-29T10:07:34 guess what: unfortunately it involves a little-tiny-bit of bfs 2011-06-29T10:08:03 are you interested? 2011-06-29T10:08:55 yeah, i can understand, i wouldnt be either 2011-06-29T10:11:34 mcstar: I am 2011-06-29T10:11:37 i was reading reader 2011-06-29T10:11:39 what are you doing? 2011-06-29T10:11:48 is this for collecting food? enemies? exploration? 2011-06-29T10:11:53 reddit* 2011-06-29T10:12:01 now it makes sense 2011-06-29T10:12:17 whatever, going from A to B 2011-06-29T10:12:34 so, first i compute the outline of water 2011-06-29T10:13:13 basically i get the 8 nbors of every water tile, add it to a set, and in the end i extract the water tiles 2011-06-29T10:14:33 *** aerique has quit IRC (Quit: ...) 2011-06-29T10:14:53 when theres no obstacle i use this basic alg.: compute the relative position of the target, and choose randomly from the directions that point toward the target 2011-06-29T10:16:23 when an ant steps into the outline of a water-block, AND the previous logic would make it step on water 2011-06-29T10:16:35 it changes behavior 2011-06-29T10:16:54 it generates a path in the outline, which brings it closer to the target 2011-06-29T10:17:21 and when i generate the possible paths in the outline i use a bit of bfs 2011-06-29T10:17:57 the ant is on a tile next to water, and the bfs computes the tiles in the outline seccesively away from it 2011-06-29T10:18:02 1,2,3,4... 2011-06-29T10:18:26 until no tile left untouched in the outline ,or similary the paths merge 2011-06-29T10:18:46 basically thats it 2011-06-29T10:18:53 (reddit distracts me :P, reading now..) 2011-06-29T10:19:45 why don't you just use a* already? :p 2011-06-29T10:19:50 naw 2011-06-29T10:20:34 this way im deling with a minimum number of squares 2011-06-29T10:20:37 dealing 2011-06-29T10:20:51 i don't know about that 2011-06-29T10:21:35 and the bfs isnt really a full-blown bfs, i exit immediately when i find the one tile i was looking for in the outline, the one closest to the target 2011-06-29T10:21:53 oh, and of course the water-blocks are separated 2011-06-29T10:22:21 so im only dealing with an outline which is the outline of a single connected pond, the one the ant tried to walk on 2011-06-29T10:23:12 that sounds both more complicated than a* and worse :P 2011-06-29T10:23:14 and when the ant is not on one of them, theres no searching at all 2011-06-29T10:23:34 i dont think its much worse 2011-06-29T10:24:02 i agree that this isnt as robust as a standard graph searching alg. 2011-06-29T10:24:33 i have converted to being a really big fan of a* now that i understand it 2011-06-29T10:24:52 hey, i told it before, i like it too, its nice 2011-06-29T10:25:02 but i didnt invent it, so ... 2011-06-29T10:25:24 *** gobang has quit IRC (Ping timeout: 255 seconds) 2011-06-29T10:25:28 i dont want to pay royalties 2011-06-29T10:25:32 i haven't seen any mention of multiple targets for a* anywhere 2011-06-29T10:25:40 pay royalties? 2011-06-29T10:25:49 yeah, to A* 2011-06-29T10:25:53 weird name though 2011-06-29T10:26:04 *** davidd has quit IRC (Quit: That's it for today) 2011-06-29T10:26:25 someone here mentioned multiple target A* before 2011-06-29T10:26:29 and it wasnt you 2011-06-29T10:32:26 antimatroid: what do you mean exactly by multiple target A*? 2011-06-29T10:33:16 mcstar: say you take an ant as a source and the set of food locations as your targets and you want to find a minimum lengthed path to a food square 2011-06-29T10:33:21 then you have multiple targets 2011-06-29T10:33:57 yes, but does the algorithm guarantees that no two ants try to step on the same tile at the same time? 2011-06-29T10:34:23 ah, sorry 2011-06-29T10:34:30 1 ant multiple food ok 2011-06-29T10:34:38 i do use multiple sources too 2011-06-29T10:34:43 but i move 1 ant per search 2011-06-29T10:34:47 im more interested in that 2011-06-29T10:34:57 multiple sources is really easy 2011-06-29T10:35:05 how do you make sure of the above? 2011-06-29T10:35:08 just add all ants locations to your queue when you start searching 2011-06-29T10:35:25 *** stocha has joined #aichallenge 2011-06-29T10:35:30 what kind of map is that one ? http://aichallengebeta.hypertriangle.com/visualizer.php?game=44778&user=199 2011-06-29T10:35:36 then when you find a shortest path to food, move the ant and start searching again 2011-06-29T10:35:48 stocha: a crap one :P 2011-06-29T10:35:50 everyone same distance to food ? so the one with an opponent that don't want to suicide wins ? 2011-06-29T10:36:08 antimatroid: it has the interesting property of how much a map can affect bot strength 2011-06-29T10:36:31 here the suicidal bots have an advantage. 2011-06-29T10:36:39 but if there all are, the game result in a draw. 2011-06-29T10:36:52 so you have to be suicidal and lucky to win :p 2011-06-29T10:36:59 stocha: thank you for hijacking antimatroid from me 2011-06-29T10:37:13 mcstar: i give him back to you sorry :) 2011-06-29T10:37:26 :D 2011-06-29T10:37:45 mcstar: I think it's quicker to use multiple sources than not to 2011-06-29T10:37:49 still a very interesting case of a map breaking most good features in bots. 2011-06-29T10:37:52 I like the new symmetric maps 2011-06-29T10:37:55 because you will find the shortest path with the one search 2011-06-29T10:38:05 smiley1983: they aren't from my generator 2011-06-29T10:38:30 i'm with _flag, they're too "closed off" 2011-06-29T10:38:32 is there a reason they're all keeping the four walls around the edge, though? I think it would be better if there were some breaks 2011-06-29T10:38:40 needs wider corridors and i like toroidalness 2011-06-29T10:38:48 antimatroid: so you add the new locations of previous ants' to a ban-list? 2011-06-29T10:38:48 i don't know about the rest. But the half distance to bot starting food is nasty :p 2011-06-29T10:38:49 ask mcleopold 2011-06-29T10:39:03 mcstar: no, i restart my search :P 2011-06-29T10:39:08 oh yes, they need some work on the actual generator to make them playable 2011-06-29T10:39:10 yeah, but thats not enough 2011-06-29T10:39:11 the nash equlibra seems to go straight at the food, and commit suicide in most cases. 2011-06-29T10:39:29 mcstar: i was thinking yesterday... you could hash information about paths for ants to food they are collecting 2011-06-29T10:39:29 you must tell the alg. that it must exclude those tiles where you already sent an ant 2011-06-29T10:40:10 will the beta use random maps soon ? (rather than a fixed set of generated one) 2011-06-29T10:40:25 then at the start of each turn, collected food that isn't expecting to be collected including your ants you're expecting to move, then after that collect ants with previous paths if a new one hasn't been found, then recollect food for what you expected to have collected but a new path was found.. 2011-06-29T10:40:31 antimatroid: there are a gigantic number of paths.. 2011-06-29T10:40:57 lol, let me read that again 2011-06-29T10:41:01 i don't mean paths between locations, i mean once you collect food, store that path 2011-06-29T10:41:18 antimatroid: what about pushing 128x128 maps as upper bound for map size ? 2011-06-29T10:41:19 and unless you find a closer food square in the time it takes to get there, keep using that path 2011-06-29T10:41:22 and don't recollect that food item 2011-06-29T10:41:40 ah 2011-06-29T10:41:42 ah 2011-06-29T10:41:52 sure i already have queues associated to ants 2011-06-29T10:42:01 *** stocha has quit IRC (Quit: Page closed) 2011-06-29T10:42:10 bye stocha 2011-06-29T10:43:04 another problem i've been thinking about is hunting down enemy ants at the end of a game for a kill 2011-06-29T10:43:24 better bots are going to try and avoid you and if the maps have long paths, they might be able to evade your vision with mazey maps 2011-06-29T10:43:39 so you really need to track information about where enemy ants out of vision might be 2011-06-29T10:43:51 i already track such information for food, that's easy 2011-06-29T10:44:34 doing it for enemies is slightly more complicated, you want to try and match up ants that you see die so you don't carry that information then flood the map slowly from the square to not-visible squares 2011-06-29T10:45:13 it will be cool to watch the chaser/evader tactics of bots here 2011-06-29T10:45:17 particularly the better bots 2011-06-29T10:45:40 this game is really awesome for ai 2011-06-29T10:45:49 its like in the beginning, searching for food, in the end you search for hiding enemies 2011-06-29T10:46:13 yeah, but you can use information 2011-06-29T10:46:21 like if you see them running down a corridor across water 2011-06-29T10:46:29 you can use that information to hunt them down 2011-06-29T10:46:34 so you want to estimate the remaining number of enemy ants? 2011-06-29T10:46:45 no, i want to estimate where ants will be 2011-06-29T10:46:51 antimatroid: i think one should store a bool[r][c] that say if an enemy can be there, and at every turn make possible_enemy[r][c] = (is visible AND an enemy is there ) OR (is not visible AND near to a cell such that possible_enemy was true the turn before) 2011-06-29T10:47:22 carlop: i will have something saying the last turn an enemy ant seen could have reached a square 2011-06-29T10:47:27 but i think this logic must be activated factoring in the number of enemies, number of own ants, and remaining turn and controlled territory 2011-06-29T10:47:59 like if an enemy ant information is outside of view, that information will be made the current turn 2011-06-29T10:48:08 and that information will "propagate" outwards every turn 2011-06-29T10:48:15 to non-visible squares 2011-06-29T10:48:18 you can associate a number whith every such turn and decrease it as the time you visited it goes by 2011-06-29T10:48:28 such turn- >such tile 2011-06-29T10:48:48 then you can do a search based off this information 2011-06-29T10:49:10 yeah, you weight the edges by this number 2011-06-29T10:49:13 basically i'm doing the same thing of you, assuming that at the beginning each un-seen cell can be an enemy 2011-06-29T10:49:13 also, if you want to avoid enemies because you have a small army, you can use information about where enemy ants are to avoid them in your searches 2011-06-29T10:49:37 of course this info can be TOTALLY wrong 2011-06-29T10:49:52 well, more importantly make information about how the minimum turns it would take an enemy ant to get to each location, which you can do with a single pass of the map 2011-06-29T10:50:18 more so if the enemy likes to coagulate 2011-06-29T10:50:30 (that sounds weird) 2011-06-29T10:52:12 does anyone except stocha have a snake bot? 2011-06-29T10:52:32 except->besides 2011-06-29T10:52:39 no, but i was thinking of how easy it would be to do success patterns 2011-06-29T10:52:50 i want a snake with attack-heads 2011-06-29T10:53:12 make a list of target squares that is your success pattern, fill those targets with ants first, then attack enemies with the rest of your ants 2011-06-29T10:53:12 and its body could swipe to collect food 2011-06-29T10:53:19 success pattern done 2011-06-29T10:53:37 so something like "come at me bro" 2011-06-29T10:53:40 written on the map 2011-06-29T10:53:55 thats a lot of ants 2011-06-29T10:53:59 or a star 2011-06-29T10:54:04 but come at me bro would be cooler :P 2011-06-29T10:54:10 ive seen a map with penises 2011-06-29T10:54:16 it was gross 2011-06-29T10:54:26 one of these new symmetric ones 2011-06-29T10:54:35 i will never get sick of poorly drawn penises 2011-06-29T10:54:39 whether intentional or not 2011-06-29T10:55:01 yeah, i wonder if the ants get distracted 2011-06-29T10:55:17 one type of four player symmetry seems to be a communist 2011-06-29T10:55:42 really? hammer and sickle-something? 2011-06-29T10:55:53 whats that called? 2011-06-29T10:55:58 by communist i mean nazi :P 2011-06-29T10:56:02 don't ask me how i got that :P 2011-06-29T10:56:07 ah, ok ive seen that 2011-06-29T10:56:19 that barbeque is pretty awesome 2011-06-29T10:56:27 im sure uve seen it 2011-06-29T10:58:53 ? 2011-06-29T10:59:17 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 5.0/20110615151330]) 2011-06-29T10:59:31 Look at this page: http://aichallengebeta.hypertriangle.com/rankings.php?page=2 Is it looks strange only for me? 2011-06-29T11:00:02 *** boegel has quit IRC (Quit: *poof!*) 2011-06-29T11:04:49 antimatroid: i meant this one, it was on imgur but i couldnt find it there 2011-06-29T11:04:51 http://failblog.files.wordpress.com/2008/08/fail-owned-bbq-design-fail.jpg 2011-06-29T11:06:03 UncleVasya: yeah, its weird that an IBM guy is using Haskell 2011-06-29T11:09:41 *** UncleVasya has quit IRC (Ping timeout: 260 seconds) 2011-06-29T11:13:54 lol how did my bot get back to third 2011-06-29T11:13:57 @rankings 2011-06-29T11:13:57 *** UncleVasya has joined #aichallenge 2011-06-29T11:14:02 antimatroid: Top 10 players: FlagCapper(86.3), sir_macelon(82.9), antimatroid(80.2), Zaphtest02(79.8), sto_greed_002(79.4), Sto03RepliCrush(77.8), Zaphus(76.7), stoha02(73.7), solifugid(73.3), stocha(72.8) 2011-06-29T11:28:52 *** |UncleVasya| <|UncleVasya|!kvirc@31.144.157.161> has joined #aichallenge 2011-06-29T11:31:51 *** UncleVasya has quit IRC (Ping timeout: 260 seconds) 2011-06-29T11:44:36 *** astevenson has joined #aichallenge 2011-06-29T11:53:49 *** gobang has joined #aichallenge 2011-06-29T11:59:25 *** Accoun has joined #aichallenge 2011-06-29T12:09:26 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-06-29T12:12:20 <_flag> Are 1v1 games going to be part of the final contest? 2011-06-29T12:14:31 <|UncleVasya|> I think it is too early for contest organizers to be sure about this. 2011-06-29T12:18:46 *** okayzed is now known as okay 2011-06-29T12:22:59 <|UncleVasya|> @rankings 2011-06-29T12:23:00 |UncleVasya|: Top 10 players: FlagCapper(85.8), sir_macelon(84.0), antimatroid(81.0), Zaphtest02(79.5), sto_greed_002(79.2), Sto03RepliCrush(78.9), UncleVasya(75.0), Zaphus(74.3), medved(74.2), rhett79(73.6) 2011-06-29T12:23:13 <|UncleVasya|> mmm, I love new maps :) 2011-06-29T12:27:23 _flag: i don't see any reason why not? 2011-06-29T12:27:31 i didn't like it when there was no spawn points 2011-06-29T12:27:37 but now i think the scoring makes sense? 2011-06-29T12:30:51 *** McLeopold1