2011-09-15T00:00:16 mleise: hi 2011-09-15T00:00:36 oh hi 2011-09-15T00:01:19 did you see what I did here http://aichallengebeta.hypertriangle.com/specification.php 2011-09-15T00:02:12 I made a few changes in your area to support it 2011-09-15T00:02:26 I wasn't sure how to go about providing support for the applet. 2011-09-15T00:02:35 mleise: I was hoping you could do that part. 2011-09-15T00:03:32 also, check out http://aichallengebeta.hypertriangle.com/test_vis.php 2011-09-15T00:04:35 mleise: gtg, cya 2011-09-15T00:04:37 *** McLeopold has left #aichallenge 2011-09-15T00:14:47 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Ping timeout: 252 seconds) 2011-09-15T00:21:40 *** ltriant has quit IRC (Read error: No route to host) 2011-09-15T00:27:18 *** dvladim has quit IRC (Ping timeout: 252 seconds) 2011-09-15T00:44:39 *** ltriant has joined #aichallenge 2011-09-15T01:56:42 *** SimonT has joined #aichallenge 2011-09-15T01:56:51 *** trololo has joined #aichallenge 2011-09-15T01:58:30 *** nplus has quit IRC (Ping timeout: 260 seconds) 2011-09-15T01:58:54 *** nplus has joined #aichallenge 2011-09-15T02:02:35 *** SimonT has quit IRC (Ping timeout: 260 seconds) 2011-09-15T02:19:24 *** dr_terrible has quit IRC (Read error: Connection reset by peer) 2011-09-15T02:19:57 *** dr_terrible has joined #aichallenge 2011-09-15T02:24:04 *** GeorgeSebastian has quit IRC (Ping timeout: 252 seconds) 2011-09-15T02:27:32 *** dr_terrible has quit IRC (Read error: Connection reset by peer) 2011-09-15T02:27:46 *** dr_terrible has joined #aichallenge 2011-09-15T02:45:08 *** Palmik has joined #aichallenge 2011-09-15T02:53:55 *** mleise has quit IRC (Remote host closed the connection) 2011-09-15T02:56:51 *** mleise has joined #aichallenge 2011-09-15T03:03:59 *** ChrisH has joined #aichallenge 2011-09-15T03:04:13 hi 2011-09-15T03:05:41 hi ChrisH 2011-09-15T03:06:13 The beta server isn't allowing new code submissions atm, know anything about it? 2011-09-15T03:07:12 No, I have no idea. 2011-09-15T03:08:05 Could be the updates from ~12 hours ago 2011-09-15T03:08:24 Ok, there's a post in the forum about it, so I guess I'll check back tomorrow. 2011-09-15T03:09:01 *** ChrisH has left #aichallenge 2011-09-15T03:11:34 *** Accoun has quit IRC (Ping timeout: 252 seconds) 2011-09-15T03:12:48 *** Accoun has joined #aichallenge 2011-09-15T03:15:18 *** Palmik has quit IRC (Read error: Connection reset by peer) 2011-09-15T03:16:12 *** Palmik has joined #aichallenge 2011-09-15T03:22:36 *** ltriant has quit IRC (Quit: Computer has gone to sleep) 2011-09-15T03:36:43 *** berak has joined #aichallenge 2011-09-15T03:38:43 *** aerique has joined #aichallenge 2011-09-15T04:20:14 does anyone know if there is a common name for the difference between types of symmetric maps? 2011-09-15T04:20:37 those where the players all have the same symmetric access to every other player and those which are just symmetric if every player plays symmetrically? 2011-09-15T04:37:18 *** boegel has joined #aichallenge 2011-09-15T05:09:44 *** TTE has joined #aichallenge 2011-09-15T05:13:15 *** mcstar has joined #aichallenge 2011-09-15T05:14:15 *** onensora has joined #aichallenge 2011-09-15T05:16:33 antimatroid: what do you mean? 2011-09-15T05:16:56 mcstar: take a 2 or 4 player game on a square grid 2011-09-15T05:17:10 the 2 player map gives each player equal access to every opponent 2011-09-15T05:17:40 *** TTE has quit IRC (Quit: Page closed) 2011-09-15T05:17:42 but the 4 player map doesn't, one of your opponents will be diagonal to you or something whereas the other two above or below and left or right 2011-09-15T05:18:08 i want to make such a distinction more generally for my game theory thesis 2011-09-15T05:18:17 i'm tempted to define the later as fair and not symmetric 2011-09-15T05:18:30 for translational symmetry its always true isnt it? 2011-09-15T05:18:46 no it's not 2011-09-15T05:19:18 for example with ants, with one ant each at the start the maximum is 5 players where each player has equal access to each of the others 2011-09-15T05:19:43 i thought it was 3, but someone on #math recently showed me you can have 5 2011-09-15T05:21:22 ah, you mean access to the exact same players? 2011-09-15T05:22:48 *** sigh has joined #aichallenge 2011-09-15T05:24:34 access to all other players 2011-09-15T05:25:00 player a can reach b,c,d and e in the same amount of turns 2011-09-15T05:25:08 that sounds impossible antimatroid 2011-09-15T05:25:32 like when there are 4 players, each has 4 opponent, but those 4 opponents cant be the same, for player a: 2xb,2xc for player b: 2xa,2xd and.... 2011-09-15T05:25:37 2 should be the only map that works for this 2011-09-15T05:25:41 2 players 2011-09-15T05:26:26 well, i think simething can be done 2011-09-15T05:26:33 you have to change the boundary conditions 2011-09-15T05:26:41 identify different edges with each other 2011-09-15T05:26:55 we are talking about "ants" 2011-09-15T05:26:59 antimatroid: still here? 2011-09-15T05:27:02 or antimatroid is 2011-09-15T05:27:06 mleise: no 2011-09-15T05:27:12 he is talking in general 2011-09-15T05:27:18 afaik 2011-09-15T05:27:22 "for example with ants, with one ant each at the start the maximum is 5 players where each player has equal access to each of the others" 2011-09-15T05:27:30 "for example" 2011-09-15T05:27:49 he talks in general, then he pulls a familiar example 2011-09-15T05:27:58 :-/ ... 5 players in ants <- i don't believe that 2011-09-15T05:28:27 how would you place 5 players such that they all have equal access to all other players 2011-09-15T05:28:55 mleise: apparently you havent seen those games? 2011-09-15T05:29:11 i'm back just reading sorry 2011-09-15T05:29:23 .a.b.c. is obviously 3 player symmetric with equal access 2011-09-15T05:29:32 5 is a bit harder, let me work an example 2011-09-15T05:29:58 lies. b has fast access to a and c, but a and c cannot reach each other 2011-09-15T05:30:12 why? 2011-09-15T05:30:23 .c.a.b 2011-09-15T05:30:33 forget the last . from antimatroid 2011-09-15T05:30:46 http://pastebin.com/kiVYWwE6 2011-09-15T05:30:47 ah now it works out 2011-09-15T05:30:53 yeah now it works, sorry 2011-09-15T05:31:00 the example i just pasted i realised that for :P 2011-09-15T05:31:09 erm, no it's not 2011-09-15T05:31:11 i fucked it up 2011-09-15T05:31:15 :) 2011-09-15T05:31:20 pastebin gonna bill you 2011-09-15T05:31:25 i wish i had the link to the image the guy made 2011-09-15T05:31:43 antimatroid: ... you are a confused prof today ^^ 2011-09-15T05:31:54 mleise: didnt you say this before? 2011-09-15T05:32:02 ill leave this channel 2011-09-15T05:32:09 i have deja vu all the time 2011-09-15T05:32:10 no i never said that before 2011-09-15T05:32:40 btw im not kidding 2011-09-15T05:32:45 its very weird 2011-09-15T05:32:55 I knew someone who had déjâ vus all the time when he was going to school 2011-09-15T05:33:38 This can appear as a real mental disease. 2011-09-15T05:34:25 mleise: i guess that diagnosis wasnt free, Dr. Doktor 2011-09-15T05:35:41 fffffff, i found the image link but the site removed the image 2011-09-15T05:35:51 hehe, I am just into this stuff. Also dreams. 2011-09-15T05:36:15 antimatroid: so you cannot proof that it works for 5 players 2011-09-15T05:36:41 prove 2011-09-15T05:37:19 i wonder why isnt a @ before my name 2011-09-15T05:39:09 it at least works if the grid isn't discrete 2011-09-15T05:39:17 Recently on YouTube: "The biggest hurricane ever sawed" 2011-09-15T05:39:57 yeah, its spelled hurrycane 2011-09-15T05:40:15 antimatroid: -.- you claimed it works for ants. now i am disappointed 2011-09-15T05:41:52 mcstar: I still imagine someone fighting the hurricane with a motor saw 2011-09-15T05:42:30 *** onensora has quit IRC () 2011-09-15T05:42:44 antimatroid: so what are the map generators doing? they set up a square lattice, and according to the number of players, they overlap another square onto this, and they scale it and rotate it, until the correct number of players fit? 2011-09-15T05:43:22 mleise: it wouldnt work, cause seawater and steel dont mix 2011-09-15T05:43:39 ah, and the smiley :) 2011-09-15T05:44:50 mcstar: my map generator does lots of things :P 2011-09-15T05:45:08 http://www.reddit.com/r/Algebra/comments/j27dt/generating_tessellations_on_a_discrete_torus/ 2011-09-15T05:45:14 that explains the tessellation maps 2011-09-15T05:45:18 ok, what i meant, how does it choose the starting positions 2011-09-15T05:45:27 aka translational 2011-09-15T05:45:31 depends on the type of symmetry 2011-09-15T05:45:46 usually i pick a location randomly and just add other players in the symmetrically equivalent locations 2011-09-15T05:46:01 although some types of symmetry have these locations closer than others, so I have a minimum starting distance 2011-09-15T05:46:22 im only considering translational symmetry 2011-09-15T05:46:55 im just saying, that you MUST have a simple square lattice, right? 2011-09-15T05:47:52 no 2011-09-15T05:48:12 http://ants.zeroviz.us/quFwc/ 2011-09-15T05:48:36 that reddit article explains the translational maps 2011-09-15T05:48:49 you might want to install the greasemonkey plugin 2011-09-15T05:49:09 http://userscripts.org/scripts/show/92758 2011-09-15T05:50:03 all trasnlational symmetry is found in the most general case on a skewed simple lattice, you just have to find the correct base-cell, with its basis vectors 2011-09-15T05:50:29 but it's not square 2011-09-15T05:50:36 it can look as hexagonal grid too, but they all come from a square lattice 2011-09-15T05:50:45 you just affinely transform it, and choose the basis 2011-09-15T05:51:15 what do you mean? 2011-09-15T05:52:21 i dont think i can be clearer tha that :S 2011-09-15T05:52:48 affine transformation - me understands 2011-09-15T05:52:50 i find parameters such that if i translate a location n times it gets back to itself 2011-09-15T05:52:51 we dont have the same background, i think in terms of crystallography 2011-09-15T05:53:06 so basically the orbit of each location has size n 2011-09-15T05:53:59 and the orbits of course partition the map 2011-09-15T05:54:50 can all maps be reduced to a skewed rectangle orbit? 2011-09-15T05:55:11 actually, it's not an orbit 2011-09-15T05:55:23 that's your name for it 2011-09-15T05:55:29 tiles maybe? 2011-09-15T05:55:33 well sure, but i was using it in the group action sense 2011-09-15T05:55:37 and in that sense it isn't 2011-09-15T05:55:50 well, maybe, but you'd be treating the locations as your group 2011-09-15T05:56:15 and that's a group if you treat it as Z_m x Z_n with the normal + operation 2011-09-15T05:56:49 and that's just weird imo 2011-09-15T05:57:25 mcstar: http://en.wikipedia.org/wiki/Group_action#Orbits_and_stabilizers 2011-09-15T05:57:37 that was what i was referring to :\ but i think i might have abused the hell out of it 2011-09-15T06:01:19 i far as ai can see, the orbit defines the primitive cell, that you will tile on a square lattice(or rectangular) 2011-09-15T06:01:33 but ill read the section before having a final conclusion :D 2011-09-15T06:02:03 well i would have called X my set of locations, not G 2011-09-15T06:02:28 it is 2011-09-15T06:02:33 G is the group, with elements g 2011-09-15T06:02:42 G.x = {g.x : g \in G} 2011-09-15T06:02:46 yeah 2011-09-15T06:02:50 that's not what i have 2011-09-15T06:07:42 so the stabilizer is the transformation that maps every x \el X to itself through the repeated action of G on X 2011-09-15T06:07:44 ? 2011-09-15T06:11:01 oh, the stabilizer isnt defined for X, only for all the x's 2011-09-15T06:11:44 in the end of the day i still think that symmetry can suck my fat foot 2011-09-15T06:12:16 im troubled with battle resolution 2011-09-15T06:12:30 ergh, rather battle solution 2011-09-15T06:13:12 quite frankly i dont see anything hidden that can be exploited to a greater advatage 2011-09-15T06:13:27 i thought of a solution, similar that revolt does 2011-09-15T06:13:48 i think that is pretty nice, but it wont work later on 2011-09-15T06:13:54 have you seen what revolt does? 2011-09-15T06:19:12 antimatroid: mleise no talky? 2011-09-15T06:19:23 thesising sorry 2011-09-15T06:19:39 thats an excuse i can understand 2011-09-15T06:22:11 trying to solve the mess McLeopold left for me to clean up :p 2011-09-15T06:23:25 *** cichociemno has joined #aichallenge 2011-09-15T06:23:43 I cannot remember the reason why I added the applet statically to the html through php magic, but he wants to add visualizers dynamically now and I don't see an immediate reason for that. I think I'll just turn that dynamic JavaScript stuff into static. 2011-09-15T06:25:25 (scrathing one's head smiley) 2011-09-15T06:27:21 mleise: my experience with the things starting with java are pretty bad 2011-09-15T06:27:42 im trying to get clojure working (actually swank) on the jvm, with no luck 2011-09-15T06:28:13 i had to look at things like ant and maven and leiningen and im in the dark 2011-09-15T06:28:32 i think ill head over #clojure 2011-09-15T06:31:03 over to* 2011-09-15T06:31:47 i think there was ibdknox who used clojure, but i havent seen him in quite a while 2011-09-15T06:31:55 anybody else using clojure here? 2011-09-15T06:36:59 mcstar: there are 3 active submissions for language 'closure' 2011-09-15T06:38:01 C++ and Java are at the top with 96 vs. 92 active submissions 2011-09-15T06:39:05 c++ is win 2011-09-15T06:45:10 mleise: clojure 2011-09-15T06:45:22 and what do you mean? 2011-09-15T06:45:39 that only means that c++ and java is easy to learn 2011-09-15T06:45:56 is->are 2011-09-15T06:46:03 (not like english) 2011-09-15T06:46:36 clojure, yes. and c++ is not all that easy. it has quite some gotchas 2011-09-15T06:47:23 antimatroid: i really hope that some day you will expand your horizon, really no disrespect, but you cant praize c++ without getting to know imho better languages 2011-09-15T06:47:26 But it is easier to switch between Java and C++ than to a functional language 2011-09-15T06:47:39 i know :P 2011-09-15T06:47:50 haskell is on the to learn list but i have other more important things to learn 2011-09-15T06:48:13 but i can and do praise c++ without knowing other languages :P 2011-09-15T06:49:00 antimatroid: certainly you can :D 2011-09-15T06:49:00 Let me see, I know Basic, Pascal/Delphi, C++, Java and JavaScript 2011-09-15T06:49:34 oh and a bit of D 2011-09-15T06:49:39 mleise: i want to use clojure for its built-in concurrency 2011-09-15T06:50:06 like foreach that splits over multiple processors? 2011-09-15T06:50:14 nah 2011-09-15T06:50:20 the agent system 2011-09-15T06:50:24 Oo 2011-09-15T06:52:05 it has a software transactinal memory, they call that, its a snapshot versioning system 2011-09-15T06:52:16 that lets agents communicate concurrently 2011-09-15T06:52:29 dont mind the multithreaded/multiprocessor side of this 2011-09-15T06:52:38 obviuosly it would run on a single thread 2011-09-15T06:52:53 what is important is the concept of this concurrency model 2011-09-15T06:53:43 it would let me treat ants as separate entities, that are really communcating, and interoperating with each other giving me a felxible way to implement a causual AI Agent model 2011-09-15T06:54:00 (to hell with my typos) 2011-09-15T06:56:38 casual -.- 2011-09-15T06:56:48 yeah 2011-09-15T06:56:51 communicating -.- 2011-09-15T06:57:02 flexible -.- 2011-09-15T06:57:10 transactional -.- 2011-09-15T06:57:19 it's -.- 2011-09-15T06:57:25 let's -.- 2011-09-15T06:57:33 hehe, yeah transactions sound nice 2011-09-15T06:58:23 so can I imagine this a bit like multiple requests to a web server, where each request sees a snapshot of the database? 2011-09-15T06:58:49 and that snapshot stays the same no matter what other requests change in the DB? 2011-09-15T06:58:55 a snapshot that was current at the time of the get request 2011-09-15T06:59:03 yeah 2011-09-15T06:59:29 that is the memory model of clojure 2011-09-15T06:59:38 and all objects are immutable 2011-09-15T06:59:38 ok the rest I think is pretty common: an event queue and worker threads 2011-09-15T06:59:52 except references, which can change WHAT do they point to 2011-09-15T06:59:54 all objects are immutable? 2011-09-15T07:00:14 otherwise, yeah 2011-09-15T07:00:16 how do you move an ant then? 2011-09-15T07:00:22 there are no locks, mutexes anything at all 2011-09-15T07:00:37 or append to a list? 2011-09-15T07:00:44 a new list is created 2011-09-15T07:00:49 holy shit 2011-09-15T07:00:59 and you can change youre reference to point to the newly created list 2011-09-15T07:01:06 your* 2011-09-15T07:01:23 holy shit is the right word 2011-09-15T07:01:27 can i have an object that contains a reference 2011-09-15T07:01:35 now imagine that every data structure works this way 2011-09-15T07:01:37 so i can change the reference inside it 2011-09-15T07:01:57 dunno 2011-09-15T07:02:01 im just a beginner 2011-09-15T07:02:12 havent actually used clojure 2011-09-15T07:02:16 I would implement a linked list then, to avoid the copies 2011-09-15T07:02:24 there are no copies 2011-09-15T07:02:30 its all efficient 2011-09-15T07:02:47 its better than linked lists 2011-09-15T07:03:18 rich hickey(the creator9 improved on lisp's main data structure, the linked list, i.e. the cons cell 2011-09-15T07:03:30 that depends on the use case. after all in clojure you recreate the list every time you add or remove 2011-09-15T07:03:40 no you dont 2011-09-15T07:03:52 but it is immutable 2011-09-15T07:03:56 yes it is 2011-09-15T07:03:59 Oo 2011-09-15T07:04:54 my list cannot be immutable and mutable at the same time 2011-09-15T07:05:08 no, its always immutable 2011-09-15T07:05:17 but you have a reference to it, that is mutable 2011-09-15T07:05:19 but then I cannot add to it !!! 2011-09-15T07:05:27 *** eashoka has joined #aichallenge 2011-09-15T07:05:40 your new list can contain as a subset the list it was created from it 2011-09-15T07:05:47 but its not a copy of the original list 2011-09-15T07:06:07 its just referenced in so to speak 2011-09-15T07:06:25 and your new list is again, immutable 2011-09-15T07:06:31 so if I add 1000 ants to a list I end up having a list referencing a sublist referencing a sublist * 1000 ?????? 2011-09-15T07:06:53 This language is pretty inefficient ^^ 2011-09-15T07:06:58 yeah, but that what a linked list is 2011-09-15T07:07:18 in lisp a linked list(a normal list) is this: 2011-09-15T07:07:20 wait, is it? 2011-09-15T07:07:24 (list a b c d) 2011-09-15T07:07:45 (cons a (cons b (cons c (cons d nil)))) 2011-09-15T07:08:01 oh yeah, it is 2011-09-15T07:08:55 so, its not really inefficient 2011-09-15T07:09:05 i mean, it not like a vector 2011-09-15T07:09:23 what if I could use a vector? 2011-09-15T07:09:23 but clojure has vectors 2011-09-15T07:09:26 or a matrix? 2011-09-15T07:09:29 sure 2011-09-15T07:09:50 what if I have an ants map and I want to set one field to 'food' ? :p 2011-09-15T07:10:24 you will have a your ants-map reference point to a new version to the 2d array, that now has at row:col a "food" 2011-09-15T07:10:47 better yet, you can create a 2d array of references 2011-09-15T07:11:01 and you can change what do they point ot 2011-09-15T07:11:02 unless, all objects are immutable 2011-09-15T07:11:20 including arrays and references inside them 2011-09-15T07:11:24 you dont create a new version of the reference-map, since you can reassign references 2011-09-15T07:11:38 i guess 2011-09-15T07:12:02 but i think this is as far as my understanding of clojure foes 2011-09-15T07:12:04 g* 2011-09-15T07:13:38 hmm ok. I hope these transactions work for you 2011-09-15T07:13:46 :D 2011-09-15T07:14:00 i hope it wont be crazy slow 2011-09-15T07:14:09 how would I do the same in say c++? 2011-09-15T07:14:22 just read the clojure source 2011-09-15T07:14:27 in a single-threaded way 2011-09-15T07:14:30 its on top of java 2011-09-15T07:14:45 this or the agent model? 2011-09-15T07:14:54 this isnt it 2011-09-15T07:15:06 this is the memory model, the agent model is something entire different 2011-09-15T07:15:10 all of it, without concurrency. maybe duplicate the complete state 2011-09-15T07:15:34 then apply functions on it, but use the original state for reference 2011-09-15T07:15:57 and as the last step move the new state over the old 2011-09-15T07:16:07 http://en.wikipedia.org/wiki/Software_transactional_memory#C.2FC.2B.2B 2011-09-15T07:16:16 :) 2011-09-15T07:17:24 at this point i consider clojure a better hype than haskell 2011-09-15T07:18:09 its increadible what is in that language, it has great features, and builds on top of lisp, and its relatively mature 2011-09-15T07:18:51 i still consider haskell as a Computer Science experiment(i dont want to hurt anyone) 2011-09-15T07:24:30 *** FireFly has joined #aichallenge 2011-09-15T07:26:38 *** delt0r__ has quit IRC (Ping timeout: 260 seconds) 2011-09-15T07:35:27 i got it working good lord 2011-09-15T07:37:12 could you blog up or mail me your CL vs Clojure experiences some day? i've been interested in it but do not see much of a point when I've already got CL experience (and I've got far too much on my plate currently to give Clojure a good ride) 2011-09-15T07:38:12 hey4 2011-09-15T07:38:14 ! 2011-09-15T07:38:31 *** delt0r__ has joined #aichallenge 2011-09-15T07:38:32 it could be THE reason to start my own blog 2011-09-15T07:38:44 sure, if i get somewhere with it i might 2011-09-15T07:58:34 now this is interesting 2011-09-15T07:58:43 my new bot .Lore. crwled back to place 39 2011-09-15T07:58:55 but i dont understand why, it has a serious bug 2011-09-15T07:59:27 basically i update an influence map badly, and after many many turns, it stops working 2011-09-15T07:59:39 the ants just wiggle around without doing much 2011-09-15T08:03:37 do you need my visual debugging tool? 2011-09-15T08:03:56 whats that? 2011-09-15T08:04:15 it is a program that waits for draw commands on a tcp socket - basically 2011-09-15T08:04:37 you could render the values of your influence map into it for each turn 2011-09-15T08:04:45 mleise: i know whats wrong, i uploaded this version, then noticed it, and right about there, mcleo closed submissions so i couldnt upload the correct one :( 2011-09-15T08:04:55 mleise: im writing them to ppm's 2011-09-15T08:05:14 and you know i have a threaded visualizer from may-june 2011-09-15T08:05:57 a threaded visualizer? 2011-09-15T08:06:23 yeah, it runs on another thread, other than the one the bot runs on 2011-09-15T08:06:37 (not multi-threaded visualizer :D) 2011-09-15T08:07:13 ok, then I guess you have what you need to get a picture 2011-09-15T08:07:22 yeah 2011-09-15T08:07:35 ppm is really easy, its ascii :D 2011-09-15T08:08:04 so it is like a svg? 2011-09-15T08:08:05 i just had to write a function that scales a floating point range to HSV 2011-09-15T08:08:13 svg is xml 2011-09-15T08:08:17 sure 2011-09-15T08:08:21 ppm is this: 2011-09-15T08:08:25 P3 2011-09-15T08:08:28 xcomment 2011-09-15T08:08:34 width heigh 2011-09-15T08:08:41 r g b r g b r g b 2011-09-15T08:08:44 and repeat that 2011-09-15T08:09:04 ah sorry, after width height theres another number that is the maximum of the rgb values 2011-09-15T08:09:12 so they range from 0 to maxrgb 2011-09-15T08:09:14 woah, that's a lot of data 2011-09-15T08:09:31 a map is below 100x100 2011-09-15T08:09:52 and who cares? its easy, and i dont want to store the pictures for long 2011-09-15T08:09:55 ok, if you always have 1 pixel per square that works 2011-09-15T08:10:09 sure, i just zoom on it in gwenview 2011-09-15T08:10:22 with scroll i change pictures 2011-09-15T08:10:22 I wanted to have the option to add arrows or text 2011-09-15T08:10:27 so i have a movie instantly 2011-09-15T08:10:56 with the visualizer i could do that, i added text, to see the ant-count of the players 2011-09-15T08:11:00 it used sdl 2011-09-15T08:11:08 ah yes i remember 2011-09-15T08:11:22 i guess it can render to images too 2011-09-15T08:11:37 probably, like cairo can render to svg 2011-09-15T08:12:58 I also added a 'map' that contains profiling info. For every stage in the process of a turn every code module is timed. So I can see what module takes the biggest time share 2011-09-15T08:18:57 *** zygon04 has quit IRC (Quit: Page closed) 2011-09-15T08:26:22 *** mathis has joined #aichallenge 2011-09-15T08:45:04 *** Apophis_ has joined #aichallenge 2011-09-15T08:45:04 *** Apophis has quit IRC (Read error: Connection reset by peer) 2011-09-15T08:47:28 mcstar: here is a picture of what it looks like http://img706.imageshack.us/img706/6932/debugpainter.png (every color is a module, and the columns are stages) 2011-09-15T08:48:58 its in german, isnt that a problem? 2011-09-15T08:49:30 ive seen that window before 2011-09-15T08:50:08 mleise: wasnt the first version of the visualizer contained in a gui like that? 2011-09-15T08:51:04 i seem to remember something like you wrote the first version in gtk-cairo? 2011-09-15T08:51:29 anyway, thats quite a detailed profiling 2011-09-15T08:51:36 since you profile each turn seőarately 2011-09-15T08:51:42 p* 2011-09-15T08:52:13 (i will clip my fingernails today) 2011-09-15T09:01:37 *** functional_spoon has joined #aichallenge 2011-09-15T09:04:02 *** trololo has quit IRC (Ping timeout: 252 seconds) 2011-09-15T09:04:24 phone 2011-09-15T09:06:01 sry i didnt hear 2011-09-15T09:19:51 mleise: i imagine there will be a huge rush to get the best names when the contest starts 2011-09-15T09:20:14 i hope mcleo will give a hint on the channel before he opens it 2011-09-15T09:20:57 its very exciting, i think my bot will have the coolest name 2011-09-15T09:25:48 *** mceier has joined #aichallenge 2011-09-15T09:30:12 i dont get this, the os is named after cats, but the wallpaper has starts and such in it 2011-09-15T09:30:14 stars* 2011-09-15T09:30:32 *** functional_spoon has quit IRC (Quit: Leaving.) 2011-09-15T09:37:18 ? 2011-09-15T09:37:20 cats 2011-09-15T09:37:25 the musical? :D 2011-09-15T09:37:46 ok just finshed a chunk of work, now i feel like slacking off a bit 2011-09-15T09:38:43 the OS 2011-09-15T09:38:46 with cats 2011-09-15T09:38:48 you know 2011-09-15T09:38:58 yeah, i guess you know :D 2011-09-15T09:39:18 good, then soon you will be able to launch your own contest delt0r 2011-09-15T09:39:31 oh yea 2011-09-15T09:39:36 with sharks and lasers 2011-09-15T09:39:41 it will be rad 2011-09-15T09:39:41 :D 2011-09-15T09:39:49 that jump over miniature motorcycles 2011-09-15T09:39:59 na.. I will finish my manager... then perhaps a entry 2011-09-15T09:40:06 and then my RTS game 2011-09-15T09:40:23 so are you really considering GA? 2011-09-15T09:40:28 oh and my fusion reactor... the world will be a better place with that 2011-09-15T09:40:32 Yep 2011-09-15T09:40:36 well GP 2011-09-15T09:40:39 dont talk me about the cluster, its no use without a proper code 2011-09-15T09:40:52 I work in evolutionary bio 2011-09-15T09:41:01 i guessed as much 2011-09-15T09:41:12 so i can even be using the cycles legit 2011-09-15T09:41:14 ive found your picture on the site 2011-09-15T09:41:21 i see you are related to dallas 2011-09-15T09:41:33 yea-- a bad one... never get round to putting up a better one 2011-09-15T09:41:37 lol 2011-09-15T09:41:41 I wish 2011-09-15T09:42:03 but my first bot does not need clusters worth of turning... 2011-09-15T09:42:07 im guessing you look older nowadays 2011-09-15T09:42:08 tuneing 2011-09-15T09:42:10 whatever 2011-09-15T09:42:23 not much.. that was taken about 4 years ago 2011-09-15T09:42:38 hm, pretty youngish i say 2011-09-15T09:42:49 and that was my daughter 2011-09-15T09:42:54 in the front. 2011-09-15T09:43:21 if you are talking about the one from the link i pasted here a few days ago 2011-09-15T09:43:57 yeah, the university stuff photo 2011-09-15T09:43:59 *** amstan has joined #aichallenge 2011-09-15T09:43:59 *** ChanServ sets mode: +o amstan 2011-09-15T09:44:14 anyway... ants! what is needed b4 launch? 2011-09-15T09:44:42 also where is the find symetric parts of the code? want to compare to my version 2011-09-15T09:44:57 * delt0r keeps reading python 2011-09-15T09:45:45 * delt0r uses grep 2011-09-15T09:49:26 mcleo mentioned cleaning up the site, and a that he forgot about a random map generation algorithm or something 2011-09-15T09:49:54 *** asfdasfsdafsdfs has joined #aichallenge 2011-09-15T09:49:55 i believe thats all what is needed, but ask am.stam 2011-09-15T09:50:02 *** asfdasfsdafsdfs has quit IRC (Client Quit) 2011-09-15T09:50:02 n 2011-09-15T09:50:51 right 2011-09-15T09:51:10 and the TrueSkill/jSkills/phySkills issues? 2011-09-15T09:51:32 McL''' is never online when i am around 2011-09-15T09:52:53 he used a php version of whatnot 2011-09-15T09:52:57 or* 2011-09-15T09:53:39 i saw you mentioned the jskills fix a couple of times, seems no one listens 2011-09-15T09:53:50 *** amstan has quit IRC (Remote host closed the connection) 2011-09-15T09:54:32 meh 2011-09-15T09:54:47 if they got something that works... fine by me 2011-09-15T10:02:08 *** janzert has quit IRC (Read error: Connection reset by peer) 2011-09-15T10:05:03 *** mogron has joined #aichallenge 2011-09-15T10:08:42 *** functional_spoon has joined #aichallenge 2011-09-15T10:10:27 *** functional_spoon has quit IRC (Client Quit) 2011-09-15T10:12:57 *** mogron has quit IRC (Quit: Page closed) 2011-09-15T10:17:04 *** aerique has quit IRC (Quit: ...) 2011-09-15T10:22:34 *** sigh has quit IRC (Remote host closed the connection) 2011-09-15T10:24:24 *** janzert has joined #aichallenge 2011-09-15T10:27:46 *** onensora has joined #aichallenge 2011-09-15T10:31:50 any word on the tcp version? 2011-09-15T10:34:28 https://github.com/berak/ants-tcp 2011-09-15T10:35:27 oh, hu. works, i think. still there could be a better player management 2011-09-15T10:35:47 is it live somewhere? 2011-09-15T10:36:00 i guess not ;( 2011-09-15T10:43:06 *** functional_spoon has joined #aichallenge 2011-09-15T11:08:19 *** McLeopold has joined #aichallenge 2011-09-15T11:08:37 delt0r: hi 2011-09-15T11:12:28 delt0r__: here I am, online... brb 2011-09-15T11:16:25 McLeopold: yay 2011-09-15T11:17:02 since i noticed you just replaced the JSkills with php skills i don't really have anything 2011-09-15T11:17:10 ask/say now 2011-09-15T11:17:19 What is outstanding? 2011-09-15T11:17:58 oh.. what is the food scheme again... tbh i am still slow reading python 2011-09-15T11:19:08 and how are nPlayer games getting sliced up : ie are there going to be say 20% 2 player 40% 4 player etc? 2011-09-15T11:26:37 *** JamesMG has quit IRC (Ping timeout: 240 seconds) 2011-09-15T11:28:40 *** boegel has quit IRC (Ping timeout: 252 seconds) 2011-09-15T11:28:46 *** functional_spoon has quit IRC (Ping timeout: 252 seconds) 2011-09-15T11:30:45 *** mathis has quit IRC (Disconnected by services) 2011-09-15T11:31:50 delt0r: n player games don't follow any requirements, other than try not to repeat games 2011-09-15T11:32:11 a seed player will be selected, then the least played map for that player will be selected. 2011-09-15T11:32:29 or for that submission, I can't remember 2011-09-15T11:32:50 The global percentages get skewed because of new entries and such 2011-09-15T11:33:31 phpskills might be a little slower, but we don't need to call out to a process and read and parse stdout, which is better from a system point of view 2011-09-15T11:33:55 before I had half the database record filled in, then waited to update the reset, so it is easier on the database as well 2011-09-15T11:34:02 mcstar: That picture was not the first visualizer. The visualizer has always been a HTML thing. It is what I wrote for my bot and it is in english. Only my bot output is in german. 2011-09-15T11:34:44 mleise: I made the visualizer dynamic for the text pages, so we can but some simple maps in the wiki, and it will translate to a nice map 2011-09-15T11:35:00 I was thinking we could write the applet tags dynamic with jQuery as well 2011-09-15T11:35:26 what about random maps? 2011-09-15T11:35:28 delt0r: a fix for JSkills would still be nice, just so we know how 2011-09-15T11:35:40 sure its a 2 line patch 2011-09-15T11:35:45 *** JamesMG has joined #aichallenge 2011-09-15T11:35:55 random as in not symmetric? 2011-09-15T11:36:00 no.. 2011-09-15T11:36:14 gtk-cairo is correct and with timing for every turn I can see if some game situations cause more work in one of the modules. But in the end it is just a product of how I defined the protocol between bot and visualization. I update all the 'pages' every turn unless they are static stuff like a grid overlay for the map 2011-09-15T11:36:34 random as in symmetric and randomly generated rather than from a very finite set of maps 2011-09-15T11:37:39 McLeopold: So the wiki would generate the 
-tags?
2011-09-15T11:38:25   also this time my manager includes a bot observer... fills in what your bot should know live... for tcp and debug
2011-09-15T11:38:47   but my visulizer is not as nice as the jnode one
2011-09-15T11:38:52   js one
2011-09-15T11:39:14   the js one runs in an applet, could you use that technique?
2011-09-15T11:39:24   I hate applets
2011-09-15T11:39:33   I have this feeling that there was a reason for not adding applets dynamically. I am not sure though.
2011-09-15T11:39:39   ;)
2011-09-15T11:39:40   well, the fact that it runs in java
2011-09-15T11:39:59   the js one runs in an applet?
2011-09-15T11:40:01   seriously ... i don't like them... both applets and WS are the worst java has to offer
2011-09-15T11:40:10   mleise: that would mean the php has to scan for the pre tags, and I hate php :)
2011-09-15T11:40:14   web start is cool delt0r
2011-09-15T11:40:51   delt0r: I want to have maps generated on a daily basis, maybe 20, but I don't want a random map for each game
2011-09-15T11:40:54   sure--if you like the fact that about 50% of the people using the link/whatever can't get anything to work
2011-09-15T11:41:06   McLeopold: Why not?
2011-09-15T11:41:18  * delt0r loves randomness and prime numbers 
2011-09-15T11:41:23   that way we can analyze results, see if we can improve on a specific map, and compare our performance on a  map to others
2011-09-15T11:41:25   mmmmm prime
2011-09-15T11:41:30   mmmmm prime
2011-09-15T11:42:11   fair enough ... would really try to stick with new maps often, or folks will scream bloody murder when new maps are added
2011-09-15T11:42:30   20 a day sounds good to me
2011-09-15T11:42:35   yea
2011-09-15T11:42:41   and I can resist screaming
2011-09-15T11:42:45  *** mathis_  has joined #aichallenge
2011-09-15T11:42:47   lol
2011-09-15T11:42:57   so, submission are turned off now...
2011-09-15T11:43:11   guess the tcp server should get more attention now?
2011-09-15T11:43:12   McLeopold: ok, I guess it will at least work on some browsers. Maybe it was just not working back when I was mixing JavaScript and applet code.
2011-09-15T11:43:49   mleise: it only needs to work on IE < 8, right?
2011-09-15T11:43:54   <=
2011-09-15T11:44:15   oh right, I would have to use Windows to do this *shudder*
2011-09-15T11:44:38   delt0r: the food scheme, after the symmetry is found, divide the food into sets that are equidistant from each start
2011-09-15T11:44:45   one day I'll set up a virtual machine
2011-09-15T11:45:06   McLeopold: More the "how many per turn" part of food.
2011-09-15T11:45:06   mleise: just get me started and I can fine tune it
2011-09-15T11:45:20   x food per y turn per player
2011-09-15T11:45:47   with ques for food that can't spawn in a turn right
2011-09-15T11:45:53   food is only spawned when the total amount can cover a food set
2011-09-15T11:45:58   that then spawns the first turn its legal
2011-09-15T11:46:13   yes, food that would spawn in an occupied location waits for it to be unoccupied
2011-09-15T11:46:51   right... so at 1 food per turn with 3 players... a set of 3 food gets spawned every 3 turns
2011-09-15T11:46:52   btw, passable means not water, unoccupied means land, I've seen some confusion on the forums
2011-09-15T11:47:04   food and ants are passable, but not unoccupied
2011-09-15T11:47:06   with the expection that a food set could be smaller than 3 from time to time
2011-09-15T11:47:24   yes
2011-09-15T11:47:38   right? so for the inital map state can an ant be on food?
2011-09-15T11:47:41   but with 3, it won't be unless 1 square is magically the same distance to all 3
2011-09-15T11:47:51   yea--sorry bad example
2011-09-15T11:47:52   no
2011-09-15T11:48:02   there is also food start
2011-09-15T11:48:22   which is 0.75% to 1.5% of the land area
2011-09-15T11:48:32   what about right next to food? does the game do a spawn before a move?
2011-09-15T11:48:39   with 2-4 being spawned within the starting view radius for each player
2011-09-15T11:48:53   oh.. so food is random
2011-09-15T11:48:59   move -> attack -> spawn
2011-09-15T11:49:21   -> then food placement
2011-09-15T11:49:41   so you can have an ant right next to a food
2011-09-15T11:49:42   yea but in the first turn... an ant could be within spawn range ie right next to food... then it could move onto the food
2011-09-15T11:49:45   which will block it
2011-09-15T11:49:50   then there is a spawn!
2011-09-15T11:50:06   the food will block
2011-09-15T11:50:17   right now... food blocks a ant in mine too... so that is "spec"
2011-09-15T11:50:22   yes
2011-09-15T11:50:28   good
2011-09-15T11:50:42   then food spawning is all that is missing...
2011-09-15T11:51:17   spec is whatever made it so we only needed one value per square :)
2011-09-15T11:51:20   finally is there a chance that we will non symmetric maps?
2011-09-15T11:51:39   no, we never did get something to analyze map performance
2011-09-15T11:52:00   right
2011-09-15T11:52:13   each map would need to know likely hood of win for each position, then that would need to be baked into the trueskill update
2011-09-15T11:52:19   it's a fairly hard problem
2011-09-15T11:52:21   I will assume that we always have symmetric
2011-09-15T11:52:36   plus, a map wouldn't be known until lots of test games were played
2011-09-15T11:52:49   i think in general this will work better anyway
2011-09-15T11:53:03   having always symmetric maps that is
2011-09-15T11:53:33   yes, symmetric works by putting each starting ant in the 8 possible symmetries and finding the position for each bot that looks the same for the others
2011-09-15T11:53:41   Oh one more question... with scoring, my understanding is that you can have a fractional score
2011-09-15T11:53:48   yes
2011-09-15T11:53:55   rounded down at the very end
2011-09-15T11:54:01   good--Ahh...
2011-09-15T11:54:03   ic
2011-09-15T11:54:27   rank is determined with the rounded down score?
2011-09-15T11:54:34   yes
2011-09-15T11:55:31   Yea, my symmetry finding code works the same way.. I assume that starting ants give me the set of possible translations too
2011-09-15T11:55:57   I wrote it to deal with multiple starting ants
2011-09-15T11:56:02   make sure you can do that as well
2011-09-15T11:56:03   same
2011-09-15T11:56:35   if the symmetry is not found, the map is considered invalid, so that is "spec" as well
2011-09-15T11:56:38   I just assume that one ant in the team has the same "view" as some other ant in every other team
2011-09-15T11:56:45   good
2011-09-15T11:56:55   same here
2011-09-15T11:57:41   for your JBotManager, if you need any options in playgame.py, I will put them in
2011-09-15T11:57:45   or the engine
2011-09-15T11:58:25   I think it is better if yours is the definition... mine is the clone --so i put in what you have
2011-09-15T11:58:44   ok
2011-09-15T11:58:47   gtg
2011-09-15T11:58:53   thxs
2011-09-15T11:58:54   later
2011-09-15T11:59:04  *** McLeopold  has left #aichallenge
2011-09-15T12:05:24   delt0r : I was following the discussion abt symmetric maps. In a 2 player map, does it symmetric around X axis, Y axis or both ?
2011-09-15T12:06:10  *** nux67  has joined #aichallenge
2011-09-15T12:29:49  *** McLeopold  has joined #aichallenge
2011-09-15T12:33:08   mleise: the js min version has an error :(
2011-09-15T12:33:33   McLeopold: did you purposely disable submissions?
2011-09-15T12:33:45   janzert: yes
2011-09-15T12:33:53   why?
2011-09-15T12:34:06   because we are ready to launch very soon
2011-09-15T12:34:11   even if we are going to launch tomorrow it doesn't seem to make sense
2011-09-15T12:34:33   why not?
2011-09-15T12:34:46   also I'm thinking we should reset all mu and sigma so we can see how the new trueskill implementation is working
2011-09-15T12:35:00   why stop people from submitting?
2011-09-15T12:35:33   I don't see us launching for a few days at least in any case
2011-09-15T12:46:19   eashoka: All possible symmetries are generated with antimatroid mapgen, so rotational, reflection and translational
2011-09-15T12:48:52   note you can also have reflection on the diagonals. ie x=y line and x=-y lines
2011-09-15T12:49:39   rotation can be 4 fold (90deg) or two fold (180 deg)
2011-09-15T12:50:28   delt0r: it's possible to have 2 valid symmetries for a map.  I'm not sure if I can define which one my algo picks
2011-09-15T12:51:22   yea--but these symmetry's are then not distinguishable
2011-09-15T12:51:33   well not from looking at the maps
2011-09-15T12:51:59   it can make a difference for the spawn code, I think
2011-09-15T12:52:46   in this case it would become distinguishable if we have the food added --or included (aka food schedule )
2011-09-15T12:53:04   but from the map itself... you can' know which one it is.
2011-09-15T12:53:27   I think in practice this would be very rare, at least on bigger maps
2011-09-15T12:53:33   Does this also means that the current maps named as "hight_map_XX" etc will no longer be used  ?
2011-09-15T12:54:49   eashoka: What map is that?
2011-09-15T12:54:56   not in my tools
2011-09-15T12:55:04   tools zip
2011-09-15T12:55:36   height_map will be used, but needs to be tweaked a little
2011-09-15T12:56:14   height_map_4 for instance. http://aichallengebeta.hypertriangle.com/map.php?map=height_map/height_map_4.map
2011-09-15T12:57:08   looks fine? Why would it not be used?
2011-09-15T12:57:21   translational symmetry
2011-09-15T12:58:37  *** cichociemno  has quit IRC (Quit: Page closed)
2011-09-15T12:58:52   delt0r & McLeopold : actually it is not needed for a map to be symmetric across all Axis isn't it ? if it is  a 2 player game, all u need is it to be symmetric along one Axis. either X or Y. if it is 4 player map, both X and Y axis.
2011-09-15T13:00:07   eashoka: or point
2011-09-15T13:01:39   how big will the maps get?
2011-09-15T13:01:43  * delt0r likes big maps
2011-09-15T13:01:56   McLeopold : u mean symmetric around a point (central point)? that would mean symmetric across all 6 axis.
2011-09-15T13:02:06   eashoka: no
2011-09-15T13:02:12   no bigger than 256, but practically 150 is almost too big for time reasons
2011-09-15T13:02:33   aww... some of my testing has been on 1000x1000 :D
2011-09-15T13:02:37   16 players
2011-09-15T13:03:11   I want to do that, but it takes way to much server time to run a game.
2011-09-15T13:03:20   true...
2011-09-15T13:03:38   well i have this cluster with 400+ cores ....
2011-09-15T13:03:42   delt04 : yea for 16 players u need all 4 axis. :-)
2011-09-15T13:04:14   eashoka: you use transnational symmetry only for >8
2011-09-15T13:04:19   ops
2011-09-15T13:04:22   translational
2011-09-15T13:04:39   no need to be bigger than a country
2011-09-15T13:05:23   ie dx*n%w=0 and dy*n%h=0
2011-09-15T13:05:40   delt0r: can we use that for the offical contest? :D
2011-09-15T13:06:00   what? transnational ?
2011-09-15T13:12:55  *** amstan  has joined #aichallenge
2011-09-15T13:12:55  *** ChanServ sets mode: +o amstan
2011-09-15T13:18:23  *** cichociemno  has joined #aichallenge
2011-09-15T13:20:51  *** trololo  has joined #aichallenge
2011-09-15T13:34:08  *** dvladim  has joined #aichallenge
2011-09-15T13:44:26   mleise: can we make ants circles and food diamonds?
2011-09-15T13:46:07   McLeopold: so, you want to launch in the weekend then?
2011-09-15T13:46:25   McLeopold: i pinged jeff last night
2011-09-15T13:46:28  *** McLeopold1  has joined #aichallenge
2011-09-15T13:46:41   McLeopold: if he's free it would be cool to meet him in the weekend and launch from toronto
2011-09-15T13:46:44  *** cichociemno  has quit IRC (Ping timeout: 252 seconds)
2011-09-15T13:46:54   McLeopold1: ^^
2011-09-15T13:46:58   if who is free?
2011-09-15T13:47:15   see scrollback
2011-09-15T13:47:25   arg, I lost connection
2011-09-15T13:48:10   okay, so jeff responded?
2011-09-15T13:48:15   no
2011-09-15T13:49:00  *** dvladim  has quit IRC (Ping timeout: 240 seconds)
2011-09-15T13:49:08  *** McLeopold  has quit IRC (Ping timeout: 260 seconds)
2011-09-15T13:49:10   jeff-a kree! launch!
2011-09-15T13:49:25   mcstar: yattack?
2011-09-15T13:49:45   lost you
2011-09-15T13:49:47   amstan: I'm thinking sep 25 for official beta, Nov 1 for official launch.
2011-09-15T13:49:50  * amstan goes for real to search a jaffa dic now
2011-09-15T13:50:04   McLeopold1: official beta? it was already official beta till now, no?
2011-09-15T13:50:22   no, I mean on the servers we will be really using
2011-09-15T13:50:40   McLeopold1: i think we're ready to do it, we just have to do a short sanity test before launching
2011-09-15T13:50:57   McLeopold1: we could do that by submitting the starter bots or something
2011-09-15T13:51:07   McLeopold1: are you willing to re-enable submissions?
2011-09-15T13:51:19   mcstar: yes, hang on
2011-09-15T13:51:25   good thx
2011-09-15T13:53:50   amstan: btw my buggy bot crawled higher than the prev. good one :S
2011-09-15T13:55:40   McLeopold : You meant Oct 1 for official Launch right .. (or does it really have to be Nov 1 ?) :-)
2011-09-15T13:55:48   oops
2011-09-15T13:55:56   haha
2011-09-15T13:56:07   i missed it too
2011-09-15T13:56:12   McLeopold1: i think we could do beta this weekend and launch next
2011-09-15T13:57:37  *** Zepp_  has joined #aichallenge
2011-09-15T13:57:46   eashoka: do you handle battles other than not entering in a no-win situation?
2011-09-15T13:57:51  *** Zepp_  has quit IRC (Client Quit)
2011-09-15T13:57:55   submissions open
2011-09-15T13:57:59   great
2011-09-15T13:59:24   yeah, it seems he does coordinate attacks
2011-09-15T13:59:41   im just wondering what is the max rank that can be reached without it
2011-09-15T14:00:42   mcstar : not clear what u meant
2011-09-15T14:01:52   eashoka: i tried my best, which part isnt clear?
2011-09-15T14:01:57   mcstar : I do some very basic battles and defence. max rank I have ever got is 3, if that answers ur qsn. :-)
2011-09-15T14:02:44   "other than not entering no-win situation"
2011-09-15T14:03:46   I looks like I'm not calculating tie ranks properly...
2011-09-15T14:04:04   this means: you do not coordinate battles, except, you try to not decrease the distance between your ant and an enemy ant, when your ant would clearly die
2011-09-15T14:04:12   for example 2 enemy vs 1 of yours
2011-09-15T14:04:15   and generalize from that
2011-09-15T14:04:39   try not to*
2011-09-15T14:05:57   oh no. I'm "trying" to do some basic battles. yea, it may still look very clumsy.. give me few more days :-)
2011-09-15T14:06:38   eashoka: you can have all the time of the world
2011-09-15T14:06:43   as far as im concerned
2011-09-15T14:06:54   damn, in* not of
2011-09-15T14:07:33   eashoka: are you willing to discuss things?
2011-09-15T14:07:42   not the battle strategy
2011-09-15T14:07:49   sure.
2011-09-15T14:08:09   im just interested in getting the max. out of the bot before focusing on battle
2011-09-15T14:08:45   I think u can reach first 15-20 without battles
2011-09-15T14:08:48   and im interested if the top bots are really there cause they focus on basic battle situations, or they do something differently(better) in the food collection business
2011-09-15T14:08:53   and suchsome
2011-09-15T14:08:57   somesuch*
2011-09-15T14:09:09   but you will need some defence to guard from those who have.
2011-09-15T14:09:12   interesting
2011-09-15T14:09:26   currently i have 3 technologies
2011-09-15T14:09:45   and defence is much easier to do than battles.
2011-09-15T14:09:54   i have 2 kinds of food collection mechanisms
2011-09-15T14:10:19   the first is a dijkstra search, i search from the food to the nearest found ant
2011-09-15T14:10:40  *** Naktibalda  has joined #aichallenge
2011-09-15T14:10:47   and i have a maximum search distance seet
2011-09-15T14:10:49   set*
2011-09-15T14:11:50   the other one is a perfect matching algorithm, which matches food with ants, the cost is the manhattan distance, then if the path is no longer than say 12, ill let the ant get the food
2011-09-15T14:12:01   i found this second one a bit better
2011-09-15T14:12:40   when the ant didnt get a food, i use an influence map for guidance
2011-09-15T14:12:56   theres a little trick there, to ensure "intelligent" motion of the ants
2011-09-15T14:13:17   hmm yea of course, to be good in any food collection calcualtion, one has to see enough of food first.
2011-09-15T14:13:48   but its mainly like other influence maps, i write a value lets say 1.0 for every ant's position, than i diffuse it for a couple of itnerations with a 9 point stencil
2011-09-15T14:13:51   so the exploration is very important first up
2011-09-15T14:14:13   yes, inflation is of key importance
2011-09-15T14:14:35   but ive seen situations, when it can be a bit disadvantageous, if the ants spreas too thin
2011-09-15T14:14:43   spread*
2011-09-15T14:15:33   you got to spread I think. little choice there, otherwise you wouldn't know what u missing :-)
2011-09-15T14:15:38   and than, i use a mask, i write out the possible attack ranges of the enemy ants to a board, and i exclude those cells from the possible next locations
2011-09-15T14:16:15   yes sure, but its not good to spread to other colonies very fast, cause you leave your explored territory unguarded
2011-09-15T14:16:47   do you handle so called "seen food"?
2011-09-15T14:16:56   yea
2011-09-15T14:17:06   i dont
2011-09-15T14:17:19   maybe i need another map for that
2011-09-15T14:17:52   another thing, is that i used a distinct water map, that used diffusion too, and i would prefereable had my ants repelled of water
2011-09-15T14:18:02   in situations when food is scarce that feature becomes important, it's effect is negligible when food is abandant
2011-09-15T14:18:05   i argue, that near water the degree of freedom of an ant decreses
2011-09-15T14:18:17   so the best is to attack from open territory
2011-09-15T14:18:48   good point with the scarce food
2011-09-15T14:18:58   i think ill implement it after all
2011-09-15T14:19:23   every little bit helps.
2011-09-15T14:19:48   so do you use similar boards like me?
2011-09-15T14:21:14   I use the existing map in Ants class to max. u can go far with that
2011-09-15T14:21:42   do you java?
2011-09-15T14:21:54   yes
2011-09-15T14:22:32   thinking of restructuring my data structures and make it bit more object oriented for ease of maintenance.
2011-09-15T14:22:57   i had an oo design, but i instead use closures now
2011-09-15T14:23:33   @later tell janzert what procedure do you use for updating the workers?
2011-09-15T14:23:33   McLeopold1: OK
2011-09-15T14:23:54   haha McLeopold1 reset the ranks
2011-09-15T14:24:04   :)
2011-09-15T14:24:35   and I'm back up to 16th!
2011-09-15T14:25:13   uh oh
2011-09-15T14:25:19   workers may be down...
2011-09-15T14:25:21   McLeopold1: he didn't seem to design the workers to be updatable
2011-09-15T14:25:24   McLeopold1: unless it's just a trivial update; update the worker tarball on the beta server then "shutdown -h now" on the worker
2011-09-15T14:25:36   McLeopold1: he just launches new ones
2011-09-15T14:25:49   amazon will bring up a new worker within about 10 minutes
2011-09-15T14:25:58   janzert: with a new id?
2011-09-15T14:26:01   yes
2011-09-15T14:26:07   new machine
2011-09-15T14:26:07   ok
2011-09-15T14:26:36   and then, it installs automatically?
2011-09-15T14:26:46   right
2011-09-15T14:26:51   sweet
2011-09-15T14:27:02   I'll be doing that in about 30 minutes
2011-09-15T14:27:12   you know the command to update the tarball?
2011-09-15T14:27:20   I did
2011-09-15T14:28:08   as the contest user; from /home/contest/aichallenge/setup run ./create_worker_archive.py ../website
2011-09-15T14:28:15   k
2011-09-15T14:29:05  *** dvladim  has joined #aichallenge
2011-09-15T14:31:37   hmm, after a ranking reset would the current pairing code still pull old ranking data?
2011-09-15T14:31:51   i.e. should the new pairing have been completely random?
2011-09-15T14:31:52   no
2011-09-15T14:31:59   yes?
2011-09-15T14:32:40   it would fall back to nearly random, but mostly submission order
2011-09-15T14:32:43   aichallenge: McLeopold epsilon * r6166a32 / (4 files in 3 dirs): support per map turn limits - http://git.io/g0JJog
2011-09-15T14:32:59   janzert: do you spot a problem?
2011-09-15T14:33:13   hmm,  we probably need it to be true random for the start of finals
2011-09-15T14:33:35   yeah...
2011-09-15T14:34:08   it's still pulling old game info, to prevent repeats
2011-09-15T14:34:09   but what triggered the question it turned out not to be true ;)
2011-09-15T14:34:39   I was misremembering the rank of a few players and thought they were previously higher than they were
2011-09-15T14:35:36   we need a backup order besides skill then
2011-09-15T14:36:01   I could do game count?
2011-09-15T14:36:22   still favors submission time
2011-09-15T14:36:44   yeah, there is no order by, so what ever order the sql returns
2011-09-15T14:37:22   the problem is that it becomes expensive to generate random numbers for every submission right?
2011-09-15T14:37:22   least game count would work, to run through the submissions in order
2011-09-15T14:38:00   so, if everyone is at the same skill, just pull players that haven't had an update yet
2011-09-15T14:38:13   I store the game count in the database now
2011-09-15T14:38:17   maybe just add a dummy field in the table to break ties and fill it with a random number?
2011-09-15T14:38:33   then before the finals we could regenerate all of them
2011-09-15T14:38:59   that would work for the first match only
2011-09-15T14:39:17   hmm, true
2011-09-15T14:39:20   the bot with the low (or high) random number would keep being selected
2011-09-15T14:39:25   game count is the way to go
2011-09-15T14:39:45   once everyone has had 2 or 3 games, skill will take over
2011-09-15T14:39:52   there's not a way to have the sql go down a different path only in the case of a tie and generate a random tie break then?
2011-09-15T14:40:20   yeah, unfortunately some people believe the first games are inordinately important
2011-09-15T14:40:34   there is, but game count ensures we move through the list and get everyone a first game before choosing bots for a second game
2011-09-15T14:41:01   game count will not exists for the first matches
2011-09-15T14:41:10   so, maybe random as well?
2011-09-15T14:41:20   submission_id is random enough for me
2011-09-15T14:41:31   if it is the third order by
2011-09-15T14:41:31   submission id is very biased
2011-09-15T14:41:40   user id is actually less biased
2011-09-15T14:42:01   milliseconds on submission id?
2011-09-15T14:42:06   at least in planetwars
2011-09-15T14:42:21   ms of submission time might work well yeah
2011-09-15T14:42:21   permute submission_id by hash?
2011-09-15T14:42:28  *** trololo  has quit IRC (Ping timeout: 252 seconds)
2011-09-15T14:42:38   or that
2011-09-15T14:42:51   seems expensive
2011-09-15T14:43:04   wait why is it expensive to generate random numbers?
2011-09-15T14:43:05   I still don't really like the static ordering it gives, but I don't think it will have a real effect
2011-09-15T14:43:48   I could generate a temp table for each matchup
2011-09-15T14:44:46   janzert: I'm going to lower the turns to 200 and reset again so we can make sure the game counts come out sort of even at the start
2011-09-15T14:45:07   ?
2011-09-15T14:45:19   why does turn limit effect game count?
2011-09-15T14:45:27   or you just want a bunch of games?
2011-09-15T14:45:34   yeah
2011-09-15T14:45:34   bunches
2011-09-15T14:45:37   that short of a limit will effect the rankings though
2011-09-15T14:45:44   meh, we are in beta
2011-09-15T14:45:44   how long do you guys estimate a bot will reach its previous ranking with the game rate/resources that will be available after the lunch?
2011-09-15T14:45:53   and then we have to try and decide if it's the new implementation or not
2011-09-15T14:46:25   mcstar: I would calculate it by # of games to drop sigma
2011-09-15T14:46:32   You can look at that data now on beta
2011-09-15T14:46:33   mcstar: depends on how fast people sign up and submit bots ;)
2011-09-15T14:46:55   i mean actual time
2011-09-15T14:46:57   janzert: the new implementation of trueskill?
2011-09-15T14:46:57   btw, the column sorting feature is buggy
2011-09-15T14:47:04   right
2011-09-15T14:47:09   what's the bug?
2011-09-15T14:47:36   it not sorting correctly ;)
2011-09-15T14:47:45   sort by ascending skill
2011-09-15T14:47:47   which column?
2011-09-15T14:47:54   the first ranking page?
2011-09-15T14:47:58   right now it's not so apparent though
2011-09-15T14:47:59   yes
2011-09-15T14:48:09   actually right now it's not apparent at all ;)
2011-09-15T14:48:25   but before the reset it would have runs of up to 10 out of order
2011-09-15T14:48:49   hmm...
2011-09-15T14:48:59   I thought I did a parseInt on those
2011-09-15T14:49:32   also I'm wondering if/afraid that since it only sorts the current page it may be more confusing than helpful?
2011-09-15T14:49:52   but that of course isn't easy to fix
2011-09-15T14:50:25   yeah, I plan to implement the jQuery pager soon
2011-09-15T14:50:32   so you get all the rankings
2011-09-15T14:50:37   ahh
2011-09-15T14:51:04   but not for game lists, they will be too long
2011-09-15T14:51:19   what does the rate column actually show now?
2011-09-15T14:51:27   games in last 24 hours
2011-09-15T14:51:28   it's not minutes between games anymore
2011-09-15T14:51:31   ok
2011-09-15T14:51:51   so I could tell if my matchmaking updates were working
2011-09-15T14:52:18   I'm open to suggestions on what that should be
2011-09-15T14:52:26   so take (34*6) / current number to get the old metric ;)
2011-09-15T14:52:32   err 24
2011-09-15T14:52:39   bah 24*60
2011-09-15T14:52:52   it doesn't really matter which it is
2011-09-15T14:53:02   I just wasn't sure what it actually was
2011-09-15T14:54:24   you know the workers aren't able to post results right now?
2011-09-15T14:54:42   or at least the one I'm looking at is saying it can't
2011-09-15T14:54:42   janzert: parseInt rounds *headslap#
2011-09-15T14:55:13   yeah, I know what's up
2011-09-15T14:55:17   ok
2011-09-15T14:55:20   I think
2011-09-15T14:55:24  *** berak  has quit IRC (Quit: ChatZilla 0.9.84 [SeaMonkey 2.0a3/20090223135443])
2011-09-15T14:55:51   just so you're aware of it before you left to go do something else :P
2011-09-15T14:55:52   janzert: parseFloat should have been used for sorting
2011-09-15T14:56:01   that makes sense
2011-09-15T14:56:03   hmm.. I'm not sure why now
2011-09-15T14:56:35   I don't remember it say "Server did not receive post" before
2011-09-15T14:56:35   oh crap
2011-09-15T14:56:45   PHPSkills is failing...
2011-09-15T14:56:49   uh oh
2011-09-15T14:56:52   :(
2011-09-15T14:57:07   it must be because the mu's and sigma's are the same
2011-09-15T14:57:13   well true skills is kinda complicated
2011-09-15T14:57:27   I'm getting Maximum execution time of 30 seconds exceeded
2011-09-15T14:57:38   and its a iterative solver --so it s bit of cpu
2011-09-15T14:57:40   ouch
2011-09-15T14:57:57   perhaps i should just patch JSkills
2011-09-15T14:58:06   it would be a lot faster than php
2011-09-15T14:58:15   delt0r: should the same patch be applied to the PHPSkills?
2011-09-15T14:58:36   I have not looked at php code since like 2001
2011-09-15T14:58:49   having it all running in php as an 'atomic' update is rather nice
2011-09-15T14:58:57   well, the php was written by moser, so it should all be the same
2011-09-15T14:58:59   so i have no idea... and tbh i would not want to look really
2011-09-15T14:59:15   ...but having it work would be better ;)
2011-09-15T14:59:16   If it really is the same thing
2011-09-15T14:59:24   well, can you post a pull request so I can see the changes?
2011-09-15T14:59:39   then basically the iterative solver won't really converge when tring to break ties
2011-09-15T15:00:13   yeah, that's probably what we are seeing now
2011-09-15T15:00:17   I am heading home now... will do it tonight (you mean JSkills right)
2011-09-15T15:00:21   yes
2011-09-15T15:01:11   ok.. well then ttyl
2011-09-15T15:01:23   maybe that's why the python version had incorrect results?
2011-09-15T15:01:52   python version seemed to have quite a few more incorrect results though
2011-09-15T15:02:20   and mostly all off in the same direction
2011-09-15T15:02:50   google says trueskill is crap on a stick
2011-09-15T15:03:14   personally i don't like the model much... but i don't have a better idea
2011-09-15T15:03:16   mcstar: do you have a link?
2011-09-15T15:03:25   heh, seems to work decently well for a fair number of games though
2011-09-15T15:03:48   http://www.gamereplays.org/community/index.php?showtopic=447220
2011-09-15T15:03:51   well, glicko would be my second choice
2011-09-15T15:04:07   glicko?
2011-09-15T15:04:09  * delt0r googles
2011-09-15T15:04:32   glicko is an improved elo but not multiplayer
2011-09-15T15:04:38  *** Allied_Envy  has quit IRC (Ping timeout: 260 seconds)
2011-09-15T15:04:47   trueskill is a modified glicko
2011-09-15T15:04:56  *** trololo  has joined #aichallenge
2011-09-15T15:05:05   how long is the code?
2011-09-15T15:05:15   a page?
2011-09-15T15:06:07   McLeopold1: what would be the preferred way of using it? what environment/language?
2011-09-15T15:09:38   mcstar: that article is what is crap, I don't think the guy understands trueskill
2011-09-15T15:10:11   he throws out numbers like "45 games" without giving beta or tau
2011-09-15T15:10:36   he uses an example sport where sigma probably needs to be reset at the start of every season
2011-09-15T15:10:44   ive read it
2011-09-15T15:10:49   you dont have to quote it
2011-09-15T15:11:10   sigma does increase for upsets
2011-09-15T15:11:25   i dont think that signma needed reset
2011-09-15T15:11:35   he meant that they play every season
2011-09-15T15:11:42   those are the games they are ranked by
2011-09-15T15:11:51   not in every season many games
2011-09-15T15:12:29   basically 1-2 matches every season, but i could misunderstand
2011-09-15T15:13:33   anyway, did you test trueskill with some weighted distribution of random numbers?
2011-09-15T15:13:47   i dont think it could be that hard to make a thorough test
2011-09-15T15:14:43   i mean you dont have to actually run the games, to have an idea if the alg. is working  properly or not
2011-09-15T15:15:46   mcstar: I can simulate 1 game where a team that is 5 times better will drop in rank below the other
2011-09-15T15:16:02   :S
2011-09-15T15:16:04   wtf
2011-09-15T15:16:12   My thought is he doesn't understand the beta values
2011-09-15T15:16:13   that does not computer
2011-09-15T15:16:15   -r
2011-09-15T15:16:22  *** Allied_Envy  has joined #aichallenge
2011-09-15T15:16:30   he said "twice as good"
2011-09-15T15:16:43   i dont really give a shat what he said
2011-09-15T15:16:54   im just asking if you tested trueskill yourself
2011-09-15T15:17:01   the beta value is the difference where 1 beats another 80% of the time
2011-09-15T15:17:04   yes
2011-09-15T15:17:13   it does work
2011-09-15T15:17:24   it works awesome
2011-09-15T15:17:40   and you didnt experience the problems(bugs) that you now seem to report?
2011-09-15T15:17:56   that isn't a problem with the math, it's with the implementation
2011-09-15T15:18:13   yeah, so what impementation did you used when testing?
2011-09-15T15:18:18   -d
2011-09-15T15:18:47  *** Accoun  has quit IRC ()
2011-09-15T15:19:01   python, which was a little of, but close enough
2011-09-15T15:19:31   what is the reason you dont want to use it?
2011-09-15T15:19:40   that little "off"?
2011-09-15T15:21:52   yes
2011-09-15T15:21:52   the problem with the python version was that it was consistently a little low with it's result
2011-09-15T15:22:12   so all the rankings would gradually drift down
2011-09-15T15:22:27   the top wasnt at 100?
2011-09-15T15:22:31   ok
2011-09-15T15:22:35   that is a problem :)
2011-09-15T15:22:43   well, now the top is at 105
2011-09-15T15:22:55   sort of fluctuates aroud 100
2011-09-15T15:23:03   i havent seen at 105 though
2011-09-15T15:23:12   the top shouldn't necessarily be at 100
2011-09-15T15:23:17   you have to mouse over the skill to see the mu
2011-09-15T15:23:28   just 95(99?) percent should be within the range 0-100
2011-09-15T15:24:03   but yes with the python implementation the average went down over time
2011-09-15T15:24:33   is this code on github? i guess it is. what folder?
2011-09-15T15:24:47   under manager
2011-09-15T15:26:43  *** delt0r__  has quit IRC (Ping timeout: 260 seconds)
2011-09-15T15:30:14   whoa that is way too complicated
2011-09-15T15:30:40  *** ChrisH  has joined #aichallenge
2011-09-15T15:31:54  *** Accoun  has joined #aichallenge
2011-09-15T15:33:03   i guess i should read the original paper if i want to understand it
2011-09-15T15:35:43  *** dvladim  has quit IRC (Ping timeout: 258 seconds)
2011-09-15T15:38:50  *** delt0r__  has joined #aichallenge
2011-09-15T15:42:41   http://www.moserware.com/2010/03/computing-your-skill.html
2011-09-15T15:42:50   that looks yummy
2011-09-15T15:43:31   i have a thing for gnuplot's wireframe graphs
2011-09-15T15:56:05  *** boegel  has joined #aichallenge
2011-09-15T15:56:27   hello, I posted a submission (in Go) about an hour ago and it hasn't been compiled yet. Is that expected?
2011-09-15T15:57:07   sort of
2011-09-15T15:57:18   the situation is hary right now
2011-09-15T15:57:27   task handing out is rather broken atm
2011-09-15T15:57:36   ok
2011-09-15T15:58:40   Reading the irc logs I see that you might be starting the "official" beta soon. Should I just wait for that, or is it likely my submission will get handled once the hairiness has been resolved?
2011-09-15T15:59:03   the submission should get handled eventually
2011-09-15T16:00:05   I'm just getting started on this contest and it's the first time I'm using and I think I'm only the second one in the beta using Go, so I'm curious if my code will compile on the game servers.
2011-09-15T16:00:12   I'm also curious which version of Go will be supported going forward.
2011-09-15T16:00:43   ChrisH: see https://github.com/aichallenge/aichallenge/issues/3
2011-09-15T16:00:45   i think youre the 3rd one
2011-09-15T16:05:02   mcstar: Thanks for the link, but although golang is mentioned in the comments, no specific versions were mentioned, and it is several months old, so I'm not sure what to take away from that.
2011-09-15T16:06:52  *** Naktibalda  has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 6.0.2/20110905175240])
2011-09-15T16:07:25  *** Allied_Envy  has quit IRC (Read error: Connection reset by peer)
2011-09-15T16:08:02   mcstar: my golang submission was just compiled successfully, so I guess that tells me what I needed to know.  thanks for your time.
2011-09-15T16:10:15   ChrisH: i guess you wanted to talk to amstan
2011-09-15T16:10:46   ChrisH: it is what it is, if you want to update it, please provide instructions
2011-09-15T16:10:48   he can probably tell you the supported version
2011-09-15T16:11:02   I'm pretty sure go tracks the latest stable version
2011-09-15T16:11:10   youcould tell him that its what comes with natty
2011-09-15T16:11:17   which may or may not be a good thing since go seems to evolve so fast
2011-09-15T16:11:24   it's not though
2011-09-15T16:11:36   at as far as I recall
2011-09-15T16:12:02   at /least/ as far as I recall
2011-09-15T16:12:14   ChrisH: if you could tell me the command to check the version i could get that for you
2011-09-15T16:12:34   amstan: I'll see what I can find...
2011-09-15T16:12:56   amstan: remember it's only installed inside the worker sandbox though
2011-09-15T16:13:10   ugh.. nvm then
2011-09-15T16:13:22   janzert: is there an easy way to get inside that?
2011-09-15T16:13:48   it's a schroot I don't remember the specific command offhand
2011-09-15T16:14:08   you can look at the sandbox code to figure it out pretty quickly though
2011-09-15T16:14:51   amstan: if you can get into the sandbox then the command would be either "5g -V", "6g -V" or "8g -V" depending on the machine's architecture.
2011-09-15T16:15:24   ChrisH: the stuff on main server reports 6g version release.2011-03-07 7666
2011-09-15T16:15:26   go is pulled from the natty version in the ppa at https://launchpad.net/~gophers/+archive/go
2011-09-15T16:15:32   ChrisH: so it might be more than that, but not less
2011-09-15T16:15:50   ok
2011-09-15T16:16:55   I am using release.r60 9481 locally. Hopefully my future submissions continue to compile.  ;)  Thanks again, everyone, for your time.
2011-09-15T16:18:01  *** ChrisH  has quit IRC (Quit: Page closed)
2011-09-15T16:21:07  *** Allied_Envy  has joined #aichallenge
2011-09-15T16:21:29   btw, the command to get in the base schroot on a worker is as root run "schroot -c aic-base"
2011-09-15T16:21:58  *** Kingpin13  has joined #aichallenge
2011-09-15T16:28:10  *** Akranis  has joined #aichallenge
2011-09-15T16:28:18  *** boegel  has quit IRC (Quit: *poof!*)
2011-09-15T16:30:26  *** eashoka  has quit IRC (Quit: Page closed)
2011-09-15T16:31:33  *** mleise  has quit IRC (Ping timeout: 260 seconds)
2011-09-15T16:32:37  *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge
2011-09-15T17:02:26  *** McLeopold1  has quit IRC (Ping timeout: 252 seconds)
2011-09-15T17:08:50  *** Palmik  has quit IRC (Remote host closed the connection)
2011-09-15T17:12:40  *** tsiki  has left #aichallenge
2011-09-15T17:23:58  *** noother  has quit IRC (Ping timeout: 260 seconds)
2011-09-15T17:26:42  *** Akranis  has quit IRC (Quit: Lämnar)
2011-09-15T17:32:58  *** trololo  has quit IRC (Ping timeout: 252 seconds)
2011-09-15T17:51:23  *** mcstar  has quit IRC (Quit: WeeChat 0.3.5)
2011-09-15T17:56:39  *** noother`  has joined #aichallenge
2011-09-15T18:05:50  *** cichociemno  has joined #aichallenge
2011-09-15T18:08:58  *** ltriant  has joined #aichallenge
2011-09-15T18:16:37  *** mceier  has quit IRC (Quit: leaving)
2011-09-15T18:18:22  *** mathis_  has quit IRC (Quit: Page closed)
2011-09-15T18:18:39  *** mathis  has joined #aichallenge
2011-09-15T18:21:34  <_flag> hello
2011-09-15T18:21:42  *** noother`  has quit IRC (Quit: Coyote finally caught me)
2011-09-15T18:26:08  *** FireFly  has quit IRC (Quit: FireFly)
2011-09-15T18:29:31  *** Kingpin13  has quit IRC (Quit: quit)
2011-09-15T19:25:16  *** ChrisH  has joined #aichallenge
2011-09-15T19:26:14  *** ChrisH  has quit IRC (Client Quit)
2011-09-15T19:38:24  *** cichociemno  has quit IRC (Quit: Page closed)
2011-09-15T19:50:22  *** mathis  has quit IRC (Quit: Page closed)
2011-09-15T20:01:15  *** kwketh  has joined #aichallenge
2011-09-15T20:35:23  *** mathis  has joined #aichallenge
2011-09-15T21:01:25  *** kwketh  has quit IRC (Quit: Page closed)
2011-09-15T21:54:30   LAME
2011-09-15T21:54:51   my banks internet banking is down till tomorrow morning, and i don't keep money in the account i get it out of
2011-09-15T22:36:52  *** smiley1983  has joined #aichallenge
2011-09-15T22:50:17  *** McLeopold  has joined #aichallenge
2011-09-15T22:52:44  *** mathis  has quit IRC (Disconnected by services)
2011-09-15T22:57:13   @later tell janzert I'm turning off the workers for now, since everything is sucking
2011-09-15T22:57:13   McLeopold: As you wish.
2011-09-15T23:20:00   aichallenge: McLeopold epsilon * r6b922df / sql/2_generate_matchup.sql : (log message trimmed)
2011-09-15T23:20:00   aichallenge: speed up of matchmaking
2011-09-15T23:20:00   aichallenge: The exclusion list is now making sure there are no repeat game for all
2011-09-15T23:20:00   aichallenge: players, not just the seed player, for the last game. The percentage
2011-09-15T23:20:00   aichallenge: of users logic is not needed since the list is larger.
2011-09-15T23:20:00   aichallenge: Positions are placed randomly because we now have symmetric maps and
2011-09-15T23:20:01   aichallenge: they don't really matter.
2011-09-15T23:20:42   delt0r: any luck on the JSkills patch?
2011-09-15T23:26:37  *** delt0r__  has quit IRC (Ping timeout: 240 seconds)
2011-09-15T23:39:22  *** delt0r__  has joined #aichallenge
2011-09-15T23:50:58  *** dvladim  has joined #aichallenge
2011-09-15T23:50:59   aichallenge: McLeopold epsilon * r1a41252 / (3 files): fix table sorting issue - http://git.io/0FxVRA