2010-11-15T00:00:25 well... this was not ready to be shared, but here it is: http://pastebin.com/yMk1i3hS 2010-11-15T00:00:59 your code is so much more pythonic :( 2010-11-15T00:01:04 my code looks horrible lol 2010-11-15T00:02:03 mine did not look that way on the first pass :) 2010-11-15T00:02:40 why does your code iterate only over the #of your fleets inbound 2010-11-15T00:02:44 and it could use a bit more work... but I have other problems to work on. 2010-11-15T00:02:53 line 49 2010-11-15T00:03:01 for t in xrange(1,M): 2010-11-15T00:03:19 M is max_turn 2010-11-15T00:03:26 junk, mine, enemy = fleets 2010-11-15T00:03:27 M = len(mine) 2010-11-15T00:03:29 nope 2010-11-15T00:03:31 M is the number of turns to calculate. the fleets arrays are just the total numbers arriving on each turn. 2010-11-15T00:03:42 nope? 2010-11-15T00:03:43 so, yes, it is the max turn. 2010-11-15T00:03:47 hmm 2010-11-15T00:03:47 see :) 2010-11-15T00:03:50 i am confused 2010-11-15T00:04:04 mine comes from: make_empty_arrays 2010-11-15T00:04:05 total # of fleets arriving? 2010-11-15T00:04:11 if there are no fleets coming in, mine is [0,0,0,0,0,0,0,0,0,0,0] 2010-11-15T00:04:16 ooh 2010-11-15T00:04:27 it's not the clearest thing in the world :) 2010-11-15T00:04:31 <---should read the full code 2010-11-15T00:04:31 nm 2010-11-15T00:04:33 noooo... its not... 2010-11-15T00:04:39 the clearest. 2010-11-15T00:05:14 I wasn't expecting to share any of it so soon... that's my excuse for the lack of explanation 2010-11-15T00:05:18 :/ 2010-11-15T00:05:26 heh 2010-11-15T00:05:37 I only knew because the it was implicit in the structure of fleets 2010-11-15T00:05:41 *guessed 2010-11-15T00:05:48 i leave a lot of comments around, cause otherwise i forget why i did something:-\ 2010-11-15T00:05:52 I don't think it is explicit in the code 2010-11-15T00:06:34 bhasker: good plan. I always end up re-solving things. This has been in my code since the start though, so I remember it now. 2010-11-15T00:06:49 heh 2010-11-15T00:07:01 I leave comments around which go out of date :D 2010-11-15T00:07:14 and i added like a ton of logging because i would keep running into bugs: -\ 2010-11-15T00:07:37 the end result of the run planet function is a list like this [(0,1,0), (1,1,13), ....] 2010-11-15T00:07:50 the first term is _always_ (0, current_owner, 0) 2010-11-15T00:08:05 btw, what is junk? 2010-11-15T00:08:06 it made some of my other code easier, where I subtract the number of ships on turn N from turn N-1 2010-11-15T00:08:06 hmm k 2010-11-15T00:08:35 junk is just so I don't think I am going to use it. the value happens to be [0], which is a list of incoming neutral fleets :) 2010-11-15T00:08:49 :P 2010-11-15T00:09:08 my other favorite in python code is to do stuff like: ign,ore,x = f(y) 2010-11-15T00:09:18 haha 2010-11-15T00:09:32 it's hard to not end up with out of date comments 2010-11-15T00:09:35 sigh: have you used pythong much? 2010-11-15T00:09:36 i try to comment sparingly 2010-11-15T00:10:00 dmj111: a bit, it's my go to language most of the time 2010-11-15T00:10:22 excellent... the next thing will make more sense then... 2010-11-15T00:10:41 I did this : ig,no,re,x = ... in some code where I was using regular expressions. 2010-11-15T00:10:45 import re 2010-11-15T00:10:53 stuff quit working in the loop. 2010-11-15T00:10:59 :) 2010-11-15T00:11:06 that's always fun 2010-11-15T00:11:32 "oops" 2010-11-15T00:11:40 ok. everyone have a good day/night. I didn't mean to stay up this long again. 2010-11-15T00:11:42 I've done something similar before, I think 2010-11-15T00:11:44 easy mistake to make 2010-11-15T00:11:45 cya 2010-11-15T00:11:50 later. 2010-11-15T00:13:00 l8r 2010-11-15T00:13:04 thanks for sharing your code 2010-11-15T00:15:24 McLeopold: all your testmaps seem to only have one "central planet"? 2010-11-15T00:16:10 *** delt0r___ has joined #aichallenge 2010-11-15T00:17:14 Top 10 players: bocsimacko(3789), Raschi(3756), sajmon(3558), asavis(3555), SUA.khb(3541), _iouri_(3538), bix0r4ever(3537), shangas(3535), george(3524), dmj111(3517) 2010-11-15T00:18:24 *** delt0r has quit IRC (Ping timeout: 276 seconds) 2010-11-15T00:55:51 hmm 2010-11-15T00:56:01 so looks like objects in python are very inefficient 2010-11-15T00:56:04 *** Mathnerd314 has quit IRC (Quit: ChatZilla 0.9.86-rdmsoft [XULRunner 1.9.2.12/20101026210630]) 2010-11-15T00:56:10 when i represent my state as a straight tuple 2010-11-15T00:56:25 the copy time goes down from 7ms to 0.68ms 2010-11-15T00:56:27 :-\ 2010-11-15T00:57:17 http://pastebin.com/BX9EnNmF 2010-11-15T00:57:23 for anyone who is interested 2010-11-15T01:00:10 i just let c++ write my copy functions 2010-11-15T01:00:12 is that a bad idea? 2010-11-15T01:01:02 *** Mjothvitnir has quit IRC (Quit: Page closed) 2010-11-15T01:01:43 antimatroid: as long as you understand what the copy function is actually doing, sure 2010-11-15T01:03:22 copy function in c++ for most structures i bit by bit copy 2010-11-15T01:03:26 unless you have pointers in there 2010-11-15T01:03:32 in which case you want to write copy constructors 2010-11-15T01:03:42 in python its not as simple 2010-11-15T01:03:54 because everything is a reference 2010-11-15T01:04:37 I find python copy simpler to get my head around, mostly because everything behaves the same way 2010-11-15T01:05:02 i guess python is just slower 2010-11-15T01:05:07 i wrote a similar c++ code 2010-11-15T01:05:14 and it runs in under 500 microseconds:- 2010-11-15T01:05:29 well, yeah... python has a lot more overhead 2010-11-15T01:05:54 the tuple version of python is close 2010-11-15T01:05:59 but the object version is very slow 2010-11-15T01:06:41 well, in the first place objects are like one big hash 2010-11-15T01:06:45 so you have that overhead 2010-11-15T01:07:01 yea i guess, but i didnt realize how much slower object copy was compared to tuples 2010-11-15T01:07:13 i didn't want to use a tuple cause then i have to index everything by number 2010-11-15T01:07:19 and thats kind of annoying to read later 2010-11-15T01:07:29 so i used an object as the container 2010-11-15T01:07:45 thinking c++ 2010-11-15T01:07:55 since in c++ all the lookup overhead is optimized away i guess 2010-11-15T01:08:07 but in python one pays the lookup overhead everytime 2010-11-15T01:08:24 so ... the object version turns out to be a lot more crappier than i expected 2010-11-15T01:08:37 python you can't assume the lookups don't change 2010-11-15T01:08:43 yea 2010-11-15T01:08:49 c++ most things are fixed at compile time 2010-11-15T01:08:56 i understand the reason now once i thought about it a bit more 2010-11-15T01:09:03 and that's why i like c++ 2010-11-15T01:09:08 cause i have no idea about any of that stuff 2010-11-15T01:09:12 c++ just works fast for me 2010-11-15T01:09:12 haha 2010-11-15T01:09:23 antimatroid :) python gives the flexibility 2010-11-15T01:09:30 you just need to understand what is the cost you pay for it 2010-11-15T01:09:31 c++ is very flexible i find 2010-11-15T01:09:36 fast really depends on what you are measuring 2010-11-15T01:09:42 there are very few things i can't do, and i expect python doesn 2010-11-15T01:09:44 t do them either 2010-11-15T01:09:53 since my experience is limited with python i don't understand the cost for everything 2010-11-15T01:10:01 sigh: any measurement is relative 2010-11-15T01:10:02 python == less development time 2010-11-15T01:10:07 so good for prototyping 2010-11-15T01:10:25 when you need raw compute speed c/c++ will always do better 2010-11-15T01:10:26 i think i'm going to learn haskell after the contest 2010-11-15T01:10:38 everyone seems to love higher level languages, and i think that'd be my one 2010-11-15T01:10:48 antimatroid c++ has its place 2010-11-15T01:10:50 and pythong its 2010-11-15T01:10:58 you just need to know what tool to use for what 2010-11-15T01:11:05 and have to do things a bit differently 2010-11-15T01:11:08 like i just learnt 2010-11-15T01:11:17 bhasker: i'm not even really a programmer, i'm more a maths and economics student than anything 2010-11-15T01:11:19 a c++ programmer will just stick the variables in a class and be done with it 2010-11-15T01:11:40 a python programmer would probably have used a tuple 2010-11-15T01:11:42 like dmj did 2010-11-15T01:11:53 <-- being a c++ programmer i stuck stuff into an object 2010-11-15T01:11:55 which makes these contests much more appealing than codejam etc. where it's more about my strategy than problem matching skills 2010-11-15T01:12:04 heh 2010-11-15T01:12:05 bhasker: I would have stuck it into an object 2010-11-15T01:12:17 only would have switched to tuple it if was too slow 2010-11-15T01:12:30 yea as i am realizing object is sloooooow 2010-11-15T01:12:32 like 10x slow 2010-11-15T01:12:35 pretty much what you did, except I would have optimised other stuff first 2010-11-15T01:12:44 yea i did the same 2010-11-15T01:12:54 like not calculating the state unless required 2010-11-15T01:12:55 except now that i got to this point where most of my time is in updateState 2010-11-15T01:13:15 well the reason i calculated them 2010-11-15T01:13:31 was it makes it easier to calculate stuff like new ships per turn per planet etc 2010-11-15T01:13:37 if i want to resolve for a lot of turns into future 2010-11-15T01:13:56 and another thing i am realizing is accessor functions are horrible in python 2010-11-15T01:14:09 the overhead of function calls for accessors is too high 2010-11-15T01:14:17 yeah, i need future states to know what i can send at each time point 2010-11-15T01:14:29 i couldn't do incremental calculation of that 2010-11-15T01:14:37 yea and its complicated to do it 2010-11-15T01:14:39 on the fly 2010-11-15T01:14:45 i would probably write a buggy version of it 2010-11-15T01:14:47 if i did it 2010-11-15T01:14:54 not really, write an indexing function, and have the code in there 2010-11-15T01:14:59 if not generated, generate, then index 2010-11-15T01:15:05 these guys are killing me: http://ai-contest.com/forum/viewtopic.php?f=3&t=1094&start=20#p7295 2010-11-15T01:16:21 *** mega1 has quit IRC (Ping timeout: 245 seconds) 2010-11-15T01:16:25 you seem to have responded appropriately 2010-11-15T01:16:48 i don't understand why people are bitching 2010-11-15T01:17:00 if your bot doesn't really take into account symmetry 2010-11-15T01:17:06 it shouldn't really do much worse 2010-11-15T01:17:15 Top 10 players: bocsimacko(3789), Raschi(3692), bix0r4ever(3562), sajmon(3550), luksian(3535), shangas(3529), dmj111(3524), Mangara(3521), george(3517), _iouri_(3490) 2010-11-15T01:17:15 and if it does, you made assumptions 2010-11-15T01:17:17 and mine seems to be doing aobut the same 2010-11-15T01:17:21 its not worse 2010-11-15T01:17:23 the map generator was not part of the spec 2010-11-15T01:17:29 bhasker: it MIGHT, unknowingly! 2010-11-15T01:17:44 bhasker: you know.. wifi might kill you too! 2010-11-15T01:17:47 amstan: bocsimackos bot isn't doing as well though 2010-11-15T01:17:54 it lost to someone with a rank over 600 before 2010-11-15T01:17:59 i'm worried about that now 2010-11-15T01:18:11 just shows that his bot isnt perfect? 2010-11-15T01:18:14 meh, rather have people playing the game, rather than the map 2010-11-15T01:18:20 http://ai-contest.com/visualizer.php?game_id=7356348 2010-11-15T01:18:22 i don't see how it reflects anything about the map itself 2010-11-15T01:18:53 he picked bad start planets 2010-11-15T01:18:58 not sure why you think he should have won that 2010-11-15T01:19:03 sigh: i agree, although it probably could ahve been made more clear that the map generator was in no way official 2010-11-15T01:19:19 i'm referring to the difference in ranking, not his choice of moves 2010-11-15T01:19:29 yeah, I agree with that 2010-11-15T01:19:35 how does that matter he lost because he made a bad move 2010-11-15T01:19:39 the ranking is immaterial 2010-11-15T01:19:42 difference in ranking is a result of choice of moves :P 2010-11-15T01:19:57 just one map on which he made a bad move 2010-11-15T01:20:09 probably something to fix in his bot's first turn logic 2010-11-15T01:20:17 to not ignore some good planets 2010-11-15T01:20:18 nearby 2010-11-15T01:20:29 i think he has some code to avoid any planet that can be contested 2010-11-15T01:20:36 and i see that his bot does that on most maps 2010-11-15T01:20:56 on turn 1 he always goes for safe planets only 2010-11-15T01:21:21 'can be contested' means, is snipable? 2010-11-15T01:21:30 kind of 2010-11-15T01:21:39 mostly planets that can be reinforced by him 2010-11-15T01:21:44 faster than they can be attacked by enemy 2010-11-15T01:21:47 so planets close by to him 2010-11-15T01:21:50 or behind his primary planet 2010-11-15T01:22:04 he never goes for contestable planets on turn one on most maps 2010-11-15T01:22:17 anything in the middle/near middle is usually offlimits for his bots 2010-11-15T01:22:53 as far as i can tell, there aren't any huge changes though, and crashes don't seem to be much of an issue near the top 2010-11-15T01:23:06 so it went reasonbly smoothly so far 2010-11-15T01:23:40 I'm glad there was a change, hopefully people will make more robust bots 2010-11-15T01:23:46 there are two changes that need making to the map generator though 2010-11-15T01:24:00 i've always been making mine robustly 2010-11-15T01:24:07 that's why i would have felt ripped off if they didn't change 2010-11-15T01:24:14 heh 2010-11-15T01:25:23 what would be interesting is if no planets were snipable on the first turn, and all middle neutrals started with ships <= its growth 2010-11-15T01:25:38 that would make for really interesting maps actually, we should get some of those on mcleopolds server 2010-11-15T01:25:58 that would be interesting 2010-11-15T01:26:16 *** hellman has joined #aichallenge 2010-11-15T01:26:35 oh god that'd be crazy aha 2010-11-15T01:27:36 alright, i'm making some of them, shouldn't be too hard 2010-11-15T01:27:37 *** amriedle has quit IRC (Quit: Lost terminal) 2010-11-15T01:28:41 bots that go for everything will probably win that map 2010-11-15T01:28:48 conservative bots will probably lose 2010-11-15T01:28:55 nope, then you'd be snipable 2010-11-15T01:29:02 not really 2010-11-15T01:29:06 yes really 2010-11-15T01:29:07 if planet cost is below growth 2010-11-15T01:29:14 then you always gain 2010-11-15T01:29:16 non-spipable means you can defend it with all your might 2010-11-15T01:29:19 by taking it 2010-11-15T01:29:23 and holding for even one turn 2010-11-15T01:29:26 so enemy has to burn more ships 2010-11-15T01:29:29 but if you take two, you don't have all your might 2010-11-15T01:29:31 than what you did to take it 2010-11-15T01:29:58 so conservative bots will be behind if they didnt take the planet on the first turn 2010-11-15T01:30:12 yes, but if you took 4 planets 2010-11-15T01:30:17 the other bot might then own you 2010-11-15T01:30:20 k let me think hold on 2010-11-15T01:30:32 if there were 20 planets of 5 growth each 2010-11-15T01:30:39 and costing 3 2010-11-15T01:30:45 i could take all of them with 4 ships 2010-11-15T01:30:47 each 2010-11-15T01:30:53 and have 20 in reserve 2010-11-15T01:30:56 now enemy counters 2010-11-15T01:31:07 he has to send at least 6 to each planet 2010-11-15T01:31:15 my growth will outpace him 2010-11-15T01:31:26 i will hold more planets than he does 2010-11-15T01:32:24 at least if i am not wrong conservative bots should lose in such situations 2010-11-15T01:33:28 argh, there is a terrible error in the map generator 2010-11-15T01:33:37 well, not breaking anything, but really stupid 2010-11-15T01:33:57 *** sigh has quit IRC (Ping timeout: 276 seconds) 2010-11-15T01:34:18 oh wait, no there's not, i'm not used to having one line of code over 2 lines 2010-11-15T01:42:12 heh 2010-11-15T01:45:30 so do you always start on a 5growth, with 100units? 2010-11-15T01:49:45 *** Rubicon-|-Cross has joined #aichallenge 2010-11-15T01:53:16 yes 2010-11-15T01:59:42 *** bhasker has quit IRC (Quit: bhasker) 2010-11-15T02:07:02 *** narnach has quit IRC (Quit: Leaving.) 2010-11-15T02:15:14 *** virdo has quit IRC (Ping timeout: 265 seconds) 2010-11-15T02:17:16 Top 10 players: bocsimacko(3783), Raschi(3635), protocolocon(3552), bix0r4ever(3549), dmj111(3525), sajmon(3525), luksian(3523), shangas(3516), _iouri_(3513), george(3502) 2010-11-15T02:35:01 *** medrimonia has joined #aichallenge 2010-11-15T02:35:10 Hello 2010-11-15T02:36:51 hi 2010-11-15T02:44:23 Does anyone have crash/timeout/invalid order issues only with the new maps too? 2010-11-15T02:44:50 add moar debugging? 2010-11-15T02:45:14 that's a very broad error 2010-11-15T02:45:29 In fact I'm debugging it on the new tcp server, but it hasn't crashed yet. 2010-11-15T03:01:05 *** jaspervdj has joined #aichallenge 2010-11-15T03:02:20 *** Arthur has joined #aichallenge 2010-11-15T03:03:58 *** Arthur has joined #aichallenge 2010-11-15T03:06:12 *** Arthur has quit IRC (Client Quit) 2010-11-15T03:17:15 Top 10 players: bix0r4ever(3820), bocsimacko(3762), protocolocon(3609), Raschi(3565), dmj111(3527), _iouri_(3524), george(3507), luksian(3506), shangas(3505), sajmon(3497) 2010-11-15T03:21:52 medrimonia: any idea what your problems are? 2010-11-15T03:22:11 i pretty much wrote the generator so i know what the maps are like 2010-11-15T03:22:40 this post http://ai-contest.com/forum/viewtopic.php?f=17&t=1094&sid=116bd16a3a09b560954fe39a1fdd5fe1&start=0 contains accurate information too 2010-11-15T03:23:16 *** Migi32 has joined #aichallenge 2010-11-15T03:26:10 *** sigh has joined #aichallenge 2010-11-15T03:29:59 I've found it! 2010-11-15T03:30:13 what was it? 2010-11-15T03:30:21 In fact I've an ID for the central planet 2010-11-15T03:30:40 there are possibly more than "one" central planets now 2010-11-15T03:30:49 *** tapwater has quit IRC (Quit: tapwater) 2010-11-15T03:30:58 And the central planet was the planete "exactly" between my home and enemy home 2010-11-15T03:31:21 that's no longer required for line symmetric maps :) 2010-11-15T03:31:30 And with the new map generator, my bot couldn't find it -> error 2010-11-15T03:31:52 how do you like the new maps? 2010-11-15T03:32:05 the gameplay is often a lot more interesting 2010-11-15T03:33:13 I find it interesting and I think it will need a couple of modification to my bot 2010-11-15T03:34:14 Because I was often making reference to the central planet 2010-11-15T03:34:50 *** shades has quit IRC (Ping timeout: 255 seconds) 2010-11-15T03:37:11 With the new generator, can we assume that the planet with the id of 0 will be one of the centrals planets antimatroid? 2010-11-15T03:40:23 medrimonia: planet 0 is always at 14,14 2010-11-15T03:40:34 actually, that will be 17,17 at some point 2010-11-15T03:40:39 Thanks a lot 2010-11-15T03:40:44 gah, don't listen to that 2010-11-15T03:41:11 That's easier to make a quick patch to avoid the crashes on official server 2010-11-15T03:41:29 it's currently at (14,14), that will change to (19,19), the maps are generated inside a circle centred around the origin then translated such that all coordinates are positive (for the visualiser) 2010-11-15T03:41:57 the old max distance was 34, which got reduced to 24 by mistake, but that will change again 2010-11-15T03:42:55 And now the minimal distance between the two homes planet is 4 right? 2010-11-15T03:43:12 http://code.google.com/p/ai-contest/source/browse/trunk/planet_wars/backend/map_generator.py 2010-11-15T03:43:12 that's the current map generator, 2 changes WILL be made, maxRadius will be 17 and a line needs to change when selecting ship counts for one lot of neutrals, but that's not really important 2010-11-15T03:43:19 it always was :) 2010-11-15T03:43:42 it was just possible for it to be less with line symmetric maps, but the maps get a lot more drawy as the players starting angles get closer 2010-11-15T03:44:08 Really? I've played on maps were the distance was 1 or 2 on the TCP server 2010-11-15T03:45:43 not possible 2010-11-15T03:45:52 they had to start on opposite sides of the centre planet 2010-11-15T03:46:07 it had to be at least 2 with min dist of 1, and min dist is officially set at 2 2010-11-15T03:46:15 so it couldn't possibly be below 4 under those restrictions 2010-11-15T03:47:42 I might be wrong. 2010-11-15T03:48:57 there is a way actually 2010-11-15T03:49:28 you could have (0,0) (0, 0.5) (0, 1.0) etc. 2010-11-15T03:49:31 but that wont happen anymore 2010-11-15T03:50:39 I think that's really good, the games will be more differents and the bots have to be less specific 2010-11-15T03:52:03 that's the idea 2010-11-15T03:52:09 the map generator was always meant to change 2010-11-15T03:52:16 possible earlier, but nothing had ever been set 2010-11-15T03:54:24 But, nothing more will change between now and the deadline? 2010-11-15T03:54:31 unlikely 2010-11-15T03:54:57 if anything, only more restrictions on what the map must meet in order to be counted as "valid" 2010-11-15T03:55:00 Cool because hotfixing during the last week would be hard 2010-11-15T03:57:55 *** shades has joined #aichallenge 2010-11-15T03:58:03 *** Rubicon-|-Cross has quit IRC (Quit: Leaving) 2010-11-15T03:59:13 Thanks for all the informations and see you later 2010-11-15T03:59:26 *** medrimonia has quit IRC () 2010-11-15T04:04:15 *** shades has quit IRC (Ping timeout: 255 seconds) 2010-11-15T04:15:34 *** wvdschel has joined #aichallenge 2010-11-15T04:15:34 *** wvdschel has joined #aichallenge 2010-11-15T04:17:17 Top 10 players: bocsimacko(3768), bix0r4ever(3680), protocolocon(3665), Mangara(3558), _iouri_(3549), Raschi(3549), luksian(3538), dmj111(3523), george(3512), shangas(3510) 2010-11-15T04:17:38 *** hellman has quit IRC (Remote host closed the connection) 2010-11-15T04:19:17 *** shades has joined #aichallenge 2010-11-15T05:09:09 *** Frontier has joined #aichallenge 2010-11-15T05:09:43 *** perror has joined #aichallenge 2010-11-15T05:16:24 *** phreeza has joined #aichallenge 2010-11-15T05:17:18 Top 10 players: bocsimacko(3762), bix0r4ever(3712), protocolocon(3630), Mangara(3573), _iouri_(3547), Raschi(3527), dmj111(3510), shangas(3508), george(3503), Xrillo(3478) 2010-11-15T05:23:32 *** Naktibalda has joined #aichallenge 2010-11-15T05:27:35 *** HangFire_ has joined #aichallenge 2010-11-15T05:28:05 Hey, can someone help me with setting up remote debug in eclipse? 2010-11-15T05:30:01 Is no one around? 2010-11-15T05:30:35 *** krokkrok has joined #aichallenge 2010-11-15T05:42:57 try #eclipse 2010-11-15T05:47:30 need new map set on TCP server 2010-11-15T05:49:47 Thanks for the tip, will check in there. 2010-11-15T05:49:55 Accoun: no one can get hold of dhartmei 2010-11-15T05:50:53 Accoun: also, the other tcp is being run off McLeopolds home connection, I think he would like it if people only ran one or two instances of their bot at a time 2010-11-15T05:51:07 it doesn't really matter yet, as you're one of the only people using it, just a heads up :) 2010-11-15T06:02:42 *** krokkrok has quit IRC (Quit: Page closed) 2010-11-15T06:10:02 antimatroid, did you try his other email address yet? 2010-11-15T06:10:31 no :P 2010-11-15T06:17:19 Top 10 players: bocsimacko(3757), protocolocon(3648), bix0r4ever(3647), medrimonia(3548), _iouri_(3540), Raschi(3512), Mangara(3509), dmj111(3499), shangas(3490), george(3488) 2010-11-15T06:21:56 *** andrei009 has joined #aichallenge 2010-11-15T06:22:18 hello 2010-11-15T06:22:51 i have a 'stupid' question ,how do i compile and run my bot if i have the c++ package ? 2010-11-15T06:23:22 I have a project file set up with codeblocks 2010-11-15T06:23:38 then i just compile the project and use the outputed .exe file for running 2010-11-15T06:23:47 http://ai-contest.com/forum/viewtopic.php?f=18&t=751 2010-11-15T06:23:52 you might enjoy that viewer too :) 2010-11-15T06:24:23 ok , i will try it now 2010-11-15T06:28:28 *** p4p4 has joined #aichallenge 2010-11-15T06:37:11 hey antimatroid , how do i add mybot.cc in bot manager? when i add i have to put the full path or what? 2010-11-15T06:37:43 you don't, you compile the project to get an .exe 2010-11-15T06:37:49 then add the .exe to jbotmanager 2010-11-15T06:37:54 with the full path name 2010-11-15T06:38:27 ok 2010-11-15T06:50:19 *** narnach has joined #aichallenge 2010-11-15T06:50:19 *** narnach is now known as narnach_mb 2010-11-15T06:53:36 *** georgeSebastian has joined #aichallenge 2010-11-15T07:00:53 *** jaspervdj has quit IRC (Ping timeout: 245 seconds) 2010-11-15T07:11:26 *** andrei009 has quit IRC (Quit: Page closed) 2010-11-15T07:15:07 *** georgeSebastian has quit IRC (Read error: Connection reset by peer) 2010-11-15T07:17:18 Top 10 players: bocsimacko(3751), bix0r4ever(3740), protocolocon(3673), luksian(3612), Mangara(3545), _iouri_(3537), Raschi(3535), dmj111(3504), shangas(3490), george(3486) 2010-11-15T07:21:31 McLeopold: almost beaten by my rage bot variant :P 2010-11-15T07:45:49 *** iNtERrUpT has joined #aichallenge 2010-11-15T07:46:59 *** sigh has quit IRC (Remote host closed the connection) 2010-11-15T07:48:13 *** sigh has joined #aichallenge 2010-11-15T08:03:47 http://www.teagames.com/games/topdog/play.php something like that would make for a cool ai contest 2010-11-15T08:04:39 *** Rubicon-|-Cross has joined #aichallenge 2010-11-15T08:17:14 *** delt0r has joined #aichallenge 2010-11-15T08:17:20 Top 10 players: bocsimacko(3776), bix0r4ever(3760), protocolocon(3661), Mangara(3564), jambachili(3556), _iouri_(3535), Raschi(3524), dmj111(3523), shangas(3489), george(3484) 2010-11-15T08:17:33 *** delt0r___ has quit IRC (Ping timeout: 260 seconds) 2010-11-15T08:18:03 *** iNtERrUpT has quit IRC (Remote host closed the connection) 2010-11-15T08:37:08 *** lavalamp_ has quit IRC (Quit: Page closed) 2010-11-15T08:42:10 *** iNtERrUpT has joined #aichallenge 2010-11-15T08:47:08 @rankings 2010-11-15T08:47:10 antimatroid: Top 10 players: bix0r4ever(3778), bocsimacko(3776), protocolocon(3644), Mangara(3591), _iouri_(3545), jambachili(3543), dmj111(3529), Raschi(3511), shangas(3490), luksian(3485) 2010-11-15T08:47:26 god damn java bot too 2010-11-15T08:57:47 *** bhasker has joined #aichallenge 2010-11-15T09:14:45 *** hellman has joined #aichallenge 2010-11-15T09:17:19 Top 10 players: bocsimacko(3780), bix0r4ever(3745), protocolocon(3635), Mangara(3599), _iouri_(3539), dmj111(3535), luksian(3524), jambachili(3523), shangas(3489), george(3479) 2010-11-15T09:24:48 antimatroid: a game like that would be cool but that implementation is retarded 2010-11-15T09:25:21 the turn rate seems to be a step function of your airspeed or something 2010-11-15T09:26:48 also, good job on the maps. totally messed up my concept for the bot i was working on but it does make them more interesting 2010-11-15T09:35:33 *** iNtERrUpT has quit IRC (Remote host closed the connection) 2010-11-15T09:52:01 *** fawek has joined #aichallenge 2010-11-15T10:06:51 *** JamesMG has joined #aichallenge 2010-11-15T10:17:21 Top 10 players: bocsimacko(3798), bix0r4ever(3689), Raschi(3640), luksian(3613), protocolocon(3595), Mangara(3565), _iouri_(3557), dmj111(3543), shangas(3494), george(3490) 2010-11-15T10:23:05 *** Ari2 has quit IRC (Remote host closed the connection) 2010-11-15T10:32:51 *** superflit has quit IRC (Quit: superflit) 2010-11-15T10:33:05 *** p4p4 has quit IRC (Ping timeout: 240 seconds) 2010-11-15T10:35:07 *** Palmik has joined #aichallenge 2010-11-15T10:49:55 *** lavalamp_ has joined #aichallenge 2010-11-15T10:52:34 *** davidd is now known as europeanswallow 2010-11-15T10:54:04 *** narnach_mb has quit IRC (Quit: Leaving.) 2010-11-15T10:57:04 *** fawek has quit IRC (Remote host closed the connection) 2010-11-15T11:01:58 *** krokkrok has joined #aichallenge 2010-11-15T11:05:55 *** rebelxt has joined #aichallenge 2010-11-15T11:08:34 *** europeanswallow is now known as davidd 2010-11-15T11:14:57 *** superflit has joined #aichallenge 2010-11-15T11:16:35 *** p4p4 has joined #aichallenge 2010-11-15T11:17:22 Top 10 players: bocsimacko(3810), bix0r4ever(3668), protocolocon(3588), Raschi(3578), _iouri_(3576), Mangara(3561), dmj111(3531), luksian(3500), shangas(3487), george(3481) 2010-11-15T11:19:31 *** McLeopold has quit IRC (Quit: Leaving) 2010-11-15T11:29:02 Oh dear my latest submission bot is terrible :'( 2010-11-15T11:35:19 *** perestrelka has quit IRC (Read error: Connection reset by peer) 2010-11-15T11:35:36 *** perestrelka has joined #aichallenge 2010-11-15T11:40:12 *** bhasker has quit IRC (Quit: bhasker) 2010-11-15T11:40:43 lavalamp_: I bet mine is worse :/ 2010-11-15T11:41:10 I was at 500 something and now I'm below 1000 2010-11-15T11:41:28 before the maps changed I was 100's and 200's :( 2010-11-15T11:41:29 it seems competition is heating up also 2010-11-15T11:41:38 Yeah 2010-11-15T11:41:49 Well my bot is crashing and I don't know why 2010-11-15T11:42:00 Works fine in all local runs and against the tcp server. 2010-11-15T11:42:16 I'm pretty sure competition didn't get 500 places tougher in the last hour XP 2010-11-15T11:42:27 d0ugal, try the new tcp server 2010-11-15T11:42:29 *** McLeopold has joined #aichallenge 2010-11-15T11:42:31 and the new maps 2010-11-15T11:42:40 amstan: ping 2010-11-15T11:42:41 @tcp 2010-11-15T11:42:42 d0ugal: tcp = (#1) http://www.benzedrine.cx/planetwars/, or (#2) http://72.44.46.68/ 2010-11-15T11:43:02 I've been using #2 2010-11-15T11:43:17 http://zeroviz.us:8080 2010-11-15T11:43:20 d0ugal: use http://zeroviz.us:8080/ 2010-11-15T11:43:31 *** zerd has quit IRC (Remote host closed the connection) 2010-11-15T11:43:36 *** zerd has joined #aichallenge 2010-11-15T11:43:56 Ahh, will do 2010-11-15T11:45:05 contestbot learn tcp as (#1) http://www.benzedrine.cx/planetwars/, or (#2) http://72.44.46.68/ (#3) http://zeroviz.us:8080 2010-11-15T11:45:05 McLeopold: I come to serve. 2010-11-15T11:45:09 @tcp 2010-11-15T11:45:10 McLeopold: tcp = (#1) http://www.benzedrine.cx/planetwars/, or (#2) http://72.44.46.68/, or (#3) (#1) http://www.benzedrine.cx/planetwars/, or (#2) http://72.44.46.68/ (#3) http://zeroviz.us:8080 2010-11-15T11:45:20 lol 2010-11-15T11:45:21 oops 2010-11-15T11:45:38 how do I fix that? 2010-11-15T11:47:54 contestbot forget tcp 3 2010-11-15T11:47:54 McLeopold: Ready to serve. 2010-11-15T11:47:57 @tcp 2010-11-15T11:47:58 McLeopold: tcp = (#1) http://www.benzedrine.cx/planetwars/, or (#2) http://72.44.46.68/ 2010-11-15T11:48:13 contestbot learn tcp as http://zeroviz.us:8080 2010-11-15T11:48:13 McLeopold: Ready to serve. 2010-11-15T11:48:16 @tcp 2010-11-15T11:48:17 McLeopold: tcp = (#1) http://www.benzedrine.cx/planetwars/, or (#2) http://72.44.46.68/, or (#3) http://zeroviz.us:8080 2010-11-15T11:49:18 *** rebelxt has left #aichallenge 2010-11-15T11:53:51 *** McLeopold has quit IRC (Quit: Leaving) 2010-11-15T11:55:07 *** krokkrok has quit IRC (Quit: Page closed) 2010-11-15T11:58:32 I am now 99% certain that an algorithm that plays PlanetWars optimally is NP-hard. 2010-11-15T12:00:39 and interestingly enough, if you were allowed to send floating point ship numbers I'd not be so sure anymore 2010-11-15T12:02:29 how would floating point ship numbers make it easier? 2010-11-15T12:03:04 (and I would like to know your reasoning) 2010-11-15T12:04:58 the same way the 0-1 knapsack problem is harder than the fractional knapsack problem 2010-11-15T12:06:56 the same way "regular" linear programming problems are easier than integer linear programming problems 2010-11-15T12:07:16 *** tapwater has joined #aichallenge 2010-11-15T12:07:42 the big problem is not where to send your ships to, but where to get them from, I think 2010-11-15T12:09:17 Oh I see you're just talking about the optimal way to capture planets? 2010-11-15T12:09:25 *order, not way 2010-11-15T12:09:37 *** Bobng has joined #aichallenge 2010-11-15T12:09:48 *** Utkarsh has quit IRC (Ping timeout: 255 seconds) 2010-11-15T12:11:04 I think my bot has a good solution to that problem (where to get ships from)... but currently it doesn't leave enough ships at home to defend so it loses to ragey on close maps :( 2010-11-15T12:12:06 *** Mistmanov has joined #aichallenge 2010-11-15T12:12:20 yea I'm not saying good strategies can't be calculated in 1 second, just that optimal ones probably can't. 2010-11-15T12:12:57 (actually I'm totally abusing the term NP-hard here) 2010-11-15T12:13:05 *** Utkarsh has joined #aichallenge 2010-11-15T12:13:17 yeah the space to search is N! where N is number of planets it's a good idea to attack 2010-11-15T12:16:14 well it's possible that you can eliminate some of those N! possibilities, maybe even enough to get polynomial time, but if this is my point: if you have an order in which you want to capture planets, from which planets will you take the ships to attack? 2010-11-15T12:16:29 *** AlliedEnvy has quit IRC (Ping timeout: 250 seconds) 2010-11-15T12:16:42 *** TheMacca has joined #aichallenge 2010-11-15T12:17:08 *** mceier has joined #aichallenge 2010-11-15T12:17:23 Top 10 players: bocsimacko(3799), bix0r4ever(3611), protocolocon(3571), _iouri_(3542), dmj111(3533), Mangara(3519), Raschi(3477), george(3474), shangas(3472), Xrillo(3448) 2010-11-15T12:17:23 ah in my view the idea order depends on which planets you have ships to draw from 2010-11-15T12:17:29 *ideal 2010-11-15T12:17:45 it most likely does 2010-11-15T12:18:39 So my bot applies one goal at a time, finding the ideal (in theory) places to get ships from to accomplish the goal. 2010-11-15T12:18:57 It builds a tree like this and sees which order allows the most goals to be accomplished. 2010-11-15T12:19:30 and it sucks atm because it's a little too good at accomplishing goals and overextends itself 2010-11-15T12:20:16 *** TheMacca has left #aichallenge 2010-11-15T12:20:40 *** AlliedEnvy has joined #aichallenge 2010-11-15T12:21:10 *** bhasker has joined #aichallenge 2010-11-15T12:22:55 lavalamp_, what do you mean with "ideal (in theory)" here? If there are multiple ways to take the same planets, how do you decide which way is ideal? 2010-11-15T12:24:06 Migi32, I define a search order function; among other things, it tries to satisfy the goal with ships that are useful for the fewest number of other goals 2010-11-15T12:27:04 but it appears that my morning changes made my bot 600 places worse so what do I know :( 2010-11-15T12:27:38 600? wow 2010-11-15T12:27:55 Are you sure you're dealing with the >23 planets on the new maps? 2010-11-15T12:28:09 it's not that 2010-11-15T12:28:14 I don't know what happened 2010-11-15T12:28:38 my bot doesn't and it's still #200 XD 2010-11-15T12:28:53 I was in the 400 and 500's on the new maps before this morning 2010-11-15T12:29:35 my old bot was in the 100 and 200's but I had to put up my new one prematurely because of the new maps 2010-11-15T12:39:12 *** bhasker_ has joined #aichallenge 2010-11-15T12:41:30 *** bhasker has quit IRC (Ping timeout: 276 seconds) 2010-11-15T12:44:55 *** krokkrok has joined #aichallenge 2010-11-15T12:45:31 *** krokkrok has quit IRC (Client Quit) 2010-11-15T12:45:49 *** krokkrok has joined #aichallenge 2010-11-15T12:47:39 howdy 2010-11-15T12:49:06 howdy 2010-11-15T12:52:12 howdy 2010-11-15T12:52:54 =) how is everyone doing here 2010-11-15T12:53:04 going mental yet? 2010-11-15T12:53:10 Yes 2010-11-15T12:53:22 I screwed up my bot and submitted it with an error 2010-11-15T12:53:39 I then didn't have time to fix it and I dropped to 4th last in the ranking board lol 2010-11-15T12:53:39 unscrew it and resubmit ;) 2010-11-15T12:53:43 not bad, still trying to iron thins out with the new rewrite 2010-11-15T12:53:52 I have, but it'll take some time to work its way back up I guess 2010-11-15T12:53:56 d0ugal: have you accounted for the change of maps? 2010-11-15T12:54:11 *** Veracious has quit IRC (Ping timeout: 255 seconds) 2010-11-15T12:54:13 ie. max planets is now 30 2010-11-15T12:54:15 antimatroid: Yeah, although I didn't have to change much 2010-11-15T12:54:27 i didn't have to change anything :P 2010-11-15T12:54:44 Actually, I didn't either - I was tweaking a few things anyway that's all 2010-11-15T12:55:24 but either way, I screwed it up and went from a fairly average rank of ~1000 to ~4400 2010-11-15T12:55:48 but now its not just crashing it'll slowly work its way back up I hope! 2010-11-15T12:56:00 well i'm still "fighting" which some other dutch guy for first place in Holland. So I decided to tweak my heuristic bot until that doubt has been lifted :P 2010-11-15T12:56:29 ha, nice 2010-11-15T12:56:34 Error323: i don't plan to let anyone else top me in australia :) 2010-11-15T12:56:47 my bot seems to be worse at the new maps, gotta tweak it I guess 2010-11-15T12:56:58 particularly dimkamindsom after his bitching about the change 2010-11-15T12:57:02 I don't think I'm confident enough to claim the UK as mine 2010-11-15T12:57:04 antimatroid: heh yeah I wanna say the same but when ppl get also in de top 50 it becomes quite hard 2010-11-15T12:57:11 I wish I could specify Scotland, that would give me more chance. 2010-11-15T12:57:15 d0udal: not with george:P 2010-11-15T12:57:53 antimatroid: Yeah :( 2010-11-15T12:57:58 Maybe I should move country. 2010-11-15T12:58:14 i'm ranked 37 atm and he is at 32 :/ 2010-11-15T12:59:19 ahhh, another person not keen on the changes 2010-11-15T12:59:28 we mustn't lose to these people :P 2010-11-15T12:59:46 Looks like Turkey is ripe for the taking; http://ai-contest.com/country_profile.php?country_id=221 2010-11-15T13:00:12 lol ye 2010-11-15T13:00:12 have the maps on the TCP server been updated as well? 2010-11-15T13:00:18 bhasker_: are you still working on it? 2010-11-15T13:00:22 Or even. http://ai-contest.com/country_profile.php?country_id=222 2010-11-15T13:00:29 i must say, i'm very surprised by the performance (or lack thereof) of asian nations in this, would that just be them not knowing of the existence of the contest? lack of interest? I can't imagine it's lack of ability 2010-11-15T13:00:54 not really 2010-11-15T13:01:03 JamesMG: nope, try here 2010-11-15T13:01:07 http://zeroviz.us:8080/ 2010-11-15T13:01:13 keep using port 995 when running your bot 2010-11-15T13:01:37 i was just tinkering with the code to see if i can write a fast max bot, in python, looks like after many tests, i need to rewrite my core state logic to use tuples instead of objects 2010-11-15T13:01:43 python objects are insanely slow to copy 2010-11-15T13:01:56 bhasker_: I found a very annoying fail in my 17.2 that would only become apparent when not running 17.0. So I switched back to 17.0 making everything i've done for the last week (alphabeta etc) obsolete :/ 2010-11-15T13:02:02 yea, that speed change before was crazy big 2010-11-15T13:02:13 bhasker_: yep, that's true 2010-11-15T13:02:23 c++ :) 2010-11-15T13:02:33 i tried rewriting in c++ 2010-11-15T13:02:37 *** yasith has joined #aichallenge 2010-11-15T13:02:37 its too much work :( 2010-11-15T13:02:42 to rewrite all of my logic in c++ 2010-11-15T13:02:46 bhasker_: but take my advise, work out a good eval function first, because otherwise its freakin useless 2010-11-15T13:02:55 hehe, i have found my rewrite painful 2010-11-15T13:02:57 and a good eval function is very timeconsuming :/ 2010-11-15T13:02:58 i will jsut steal yours:p 2010-11-15T13:03:05 mine really really sux 2010-11-15T13:03:06 *** Veracious has joined #aichallenge 2010-11-15T13:03:11 heh i just want to do it for fun 2010-11-15T13:03:14 not really going to submit it 2010-11-15T13:03:20 so i will tinker and see 2010-11-15T13:03:25 you might as well submit at the end :P 2010-11-15T13:03:26 once i rewrite the state handling code 2010-11-15T13:03:26 thanks antimatroid 2010-11-15T13:03:40 which requires substantial changes, since i index by name everything 2010-11-15T13:03:46 i will have to change the code all over 2010-11-15T13:03:52 i don't expect to win at this point, but top 10 would satisfy me 2010-11-15T13:03:53 to index state by number 2010-11-15T13:03:59 *** medrimonia has joined #aichallenge 2010-11-15T13:04:17 Hi everyone 2010-11-15T13:04:18 it will make the code rather unreadable but thats the price one pays 2010-11-15T13:04:20 imo a gamestate is nothing more then all_planets, all_fleets right? 2010-11-15T13:04:42 well for me it is anyway 2010-11-15T13:05:28 *** krokkrok has quit IRC (Quit: Page closed) 2010-11-15T13:05:53 my gamestate has turn, distances (pre calculated), bestPlanets, planets, futureStates, numeroud count vectors, a timer and an output file stream for debugging :) 2010-11-15T13:06:20 my state is basically for each planet 2010-11-15T13:06:30 the ofstream is just a pointer so I can have the same thing copied all over the place 2010-11-15T13:06:47 mmm 2010-11-15T13:06:55 an array of state objects containing ( 4-5 numbers, #ships,#owner,#myShips on that turn, #enemyShips,#origShips before battle resolution) 2010-11-15T13:06:58 i did that orginally, but got sick of writing planets[source].futurestates[t] etc. 2010-11-15T13:07:05 sounds like a lot of extra obsolete stuff 2010-11-15T13:07:14 i basically calculate maxDistance*2+1 states for each planet 2010-11-15T13:07:23 and use that to decide how many ships i need to send where etc 2010-11-15T13:07:26 and how many are available 2010-11-15T13:07:29 yeah, i do one more state than that 2010-11-15T13:07:44 so for a typical map it can be 2010-11-15T13:07:53 roughly 50*23 state objects 2010-11-15T13:07:55 roughly 2010-11-15T13:07:58 i have a very different approach lol xD 2010-11-15T13:08:04 most maps have a turn distance of 23-24 max 2010-11-15T13:08:23 yeah, the new generator has max dist of 24 2010-11-15T13:08:28 but that will change to 34 2010-11-15T13:13:59 *** Utkarsh has quit IRC (Ping timeout: 255 seconds) 2010-11-15T13:13:59 *** _Utkarsh <_Utkarsh!~admin@117.201.83.221> has joined #aichallenge 2010-11-15T13:16:30 ah k 2010-11-15T13:17:22 Top 10 players: bocsimacko(3779), bix0r4ever(3697), protocolocon(3616), _iouri_(3551), dmj111(3548), Mangara(3545), Raschi(3492), luksian(3476), shangas(3476), george(3476) 2010-11-15T13:17:51 I do get frustrated everytime that top 10 comes up without me in it :/ 2010-11-15T13:19:06 i'm working on that^ 2010-11-15T13:21:41 who isn't :P 2010-11-15T13:23:14 * Zannick 2010-11-15T13:23:45 I do get frustrated everytime that I "improve" my bot and it gets ranked lower 2010-11-15T13:23:59 hehe yeah 2010-11-15T13:24:56 I started rewriting almost all of my bot a month ago and I've just succeed to get one better than the old one 2010-11-15T13:28:18 *** jaspervdj has joined #aichallenge 2010-11-15T13:28:38 hehe 2010-11-15T13:28:45 medrimonia: well your beating mine 2010-11-15T13:28:49 so thats not good :P 2010-11-15T13:31:06 Error323: Well until last Wednesday, this version of the bot wasn't even ranked in the top 100 2010-11-15T13:31:26 I've just correct some (very bad) mistakes 2010-11-15T13:31:32 medrimonia: well that just bad luck, it depends on your way up 2010-11-15T13:31:41 And now it seems that it's a lot better 2010-11-15T13:32:10 I don't think that I'll keep ranked like that 2010-11-15T13:32:28 I will go down soon I think 2010-11-15T13:32:43 ohh 2010-11-15T13:32:47 you just submitted 2010-11-15T13:32:50 then prolly no 2010-11-15T13:33:40 It has played less than fifty games so I think that it's not stabilized yet 2010-11-15T13:33:48 on tcp you need about 500 games for a stable rank faik 2010-11-15T13:33:50 afaik* 2010-11-15T13:33:55 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716]) 2010-11-15T13:33:59 i figure it won't be much different on official 2010-11-15T13:35:44 Yes maybe a little bit less, but the thing that is really good now it's that you keep fighting against your levels bot when you submit a new bot 2010-11-15T13:36:07 yup 2010-11-15T13:36:17 Error323: I wouldn't be surprised if you need more on official, given the much higher number of people 2010-11-15T13:36:41 And with the numbers of games per minute being rarely low, it's not as long as before to get a good rank 2010-11-15T13:36:45 That and it'll only be stable if everybody stops changing and re-uploading 2010-11-15T13:37:16 yes 2010-11-15T13:37:47 I think that the ranking begin to "stabilize" a little bit after something like ten or fifteen games. 2010-11-15T13:37:59 nah 2010-11-15T13:38:00 but It won't be accurate 2010-11-15T13:38:04 not necessarily 2010-11-15T13:38:20 you need much more 2010-11-15T13:38:38 If you're something like 300 after 10 games you will hardly be in top 10 after 300 games 2010-11-15T13:39:12 well it can happen 2010-11-15T13:39:29 and is also highly dependent on the enemies ur facing 2010-11-15T13:39:39 and of the maps 2010-11-15T13:39:56 yep 2010-11-15T13:39:59 I know that my bot can be very weak on some maps 2010-11-15T13:40:00 *** yasith has quit IRC (Ping timeout: 272 seconds) 2010-11-15T13:41:38 well even ppl with over 300 games still bounce +/- 10 ranks 2010-11-15T13:41:56 but thats also because of resubmissions indd 2010-11-15T13:42:05 of other 2010-11-15T13:42:07 s 2010-11-15T13:42:53 *** bhasker has joined #aichallenge 2010-11-15T13:43:43 If you stop uploading bot, you will loose rankings usually 2010-11-15T13:45:35 *** bhasker_ has quit IRC (Ping timeout: 264 seconds) 2010-11-15T13:47:01 *** fawek has joined #aichallenge 2010-11-15T13:50:13 I know that my bot can be very weak on all maps :/ 2010-11-15T13:53:37 *** _Utkarsh is now known as Utkarsh 2010-11-15T13:56:20 *** Cold-Phoenix has joined #aichallenge 2010-11-15T13:59:56 *** bhasker_ has joined #aichallenge 2010-11-15T14:01:02 *** perror has quit IRC (Quit: Bye all !) 2010-11-15T14:01:14 *** Utkarsh has quit IRC (Ping timeout: 255 seconds) 2010-11-15T14:02:22 *** choas has joined #aichallenge 2010-11-15T14:02:36 *** bhasker has quit IRC (Ping timeout: 240 seconds) 2010-11-15T14:04:56 *** McLeopold has joined #aichallenge 2010-11-15T14:08:58 *** Utkarsh has joined #aichallenge 2010-11-15T14:11:10 *** McLeopold_ has joined #aichallenge 2010-11-15T14:12:07 *** McLeopold_ has quit IRC (Client Quit) 2010-11-15T14:15:06 *** McLeopold has quit IRC (Ping timeout: 276 seconds) 2010-11-15T14:17:23 Top 10 players: bocsimacko(3786), bix0r4ever(3700), protocolocon(3602), Mangara(3557), _iouri_(3550), dmj111(3546), george(3481), shangas(3481), Xrillo(3474), medrimonia(3459) 2010-11-15T14:32:43 *** Cold-Phoenix has quit IRC () 2010-11-15T14:36:41 *** bhasker has joined #aichallenge 2010-11-15T14:38:49 *** bhasker_ has quit IRC (Read error: Operation timed out) 2010-11-15T14:44:36 *** McLeopold has joined #aichallenge 2010-11-15T14:45:43 @rankings 2010-11-15T14:45:44 McLeopold: Top 10 players: bix0r4ever(3807), bocsimacko(3797), protocolocon(3622), Mangara(3562), dmj111(3546), _iouri_(3533), george(3487), shangas(3482), Xrillo(3476), jambachili(3469) 2010-11-15T14:46:02 shake up in the top ten 2010-11-15T14:47:05 *** Frontier has quit IRC (Ping timeout: 240 seconds) 2010-11-15T14:48:38 wow, someone beat mega1 2010-11-15T14:49:24 http://zeroviz.us:8080/canvas?game_id=6749 2010-11-15T14:49:30 pretty funny game 2010-11-15T14:50:10 *** amriedle has joined #aichallenge 2010-11-15T14:50:31 lol 2010-11-15T14:50:45 quite a map 2010-11-15T14:54:44 *** wvdschel has quit IRC (Ping timeout: 272 seconds) 2010-11-15T14:56:58 that map draw a lot 2010-11-15T15:05:35 *** wvdschel has joined #aichallenge 2010-11-15T15:10:37 *** hellman has quit IRC (Remote host closed the connection) 2010-11-15T15:12:16 I hate it when a calculation goes a little bit wrong. http://ai-contest.com/visualizer.php?game_id=7450942 2010-11-15T15:13:10 haha 2010-11-15T15:14:42 I like to toy with them and keep 'em pinned down :P 2010-11-15T15:17:03 funny game 2010-11-15T15:17:25 Top 10 players: bix0r4ever(3809), bocsimacko(3803), protocolocon(3604), dmj111(3555), Mangara(3548), _iouri_(3544), george(3491), shangas(3486), Xrillo(3483), luksian(3464) 2010-11-15T15:18:18 hah nice game 2010-11-15T15:18:19 http://ai-contest.com/visualizer.php?game_id=7444860 2010-11-15T15:18:35 my bot surprises me sometimes 2010-11-15T15:19:03 another one http://ai-contest.com/visualizer.php?game_id=7442521 2010-11-15T15:22:38 *** Accoun has quit IRC () 2010-11-15T15:26:50 *** bhasker_ has joined #aichallenge 2010-11-15T15:29:08 *** bhasker has quit IRC (Ping timeout: 265 seconds) 2010-11-15T15:29:16 bhasker_: awesome 2010-11-15T15:29:31 yea that game with mistmanov was pretty neat turnaround 2010-11-15T15:29:52 and the fact that mistmanovx is a fellow countrymen here and you just made him drop points getting me back on helps =) 2010-11-15T15:30:17 very nice game indeed 2010-11-15T15:30:20 *** amriedle has quit IRC (Quit: Changing server) 2010-11-15T15:30:21 complex 2010-11-15T15:30:27 -_- 2010-11-15T15:30:33 hahaha oops 2010-11-15T15:30:36 hey Mistmanov ^_^ 2010-11-15T15:30:47 good evening ;) 2010-11-15T15:31:06 good evening ;) 2010-11-15T15:31:21 soooo ur gonna work again on it :P 2010-11-15T15:31:22 damn 2010-11-15T15:31:28 that gives us both more work :P 2010-11-15T15:31:41 nah, I dunno :p 2010-11-15T15:31:49 have already made a lot of changes 2010-11-15T15:32:01 but without the final piece of the puzzle 2010-11-15T15:32:11 the new bot is actually worst than the one I already uploaded -_- 2010-11-15T15:32:20 yeah I had that happened too :/ 2010-11-15T15:32:23 and I don't feel like making the last piece 2010-11-15T15:32:25 *** McLeopold_ has joined #aichallenge 2010-11-15T15:32:29 this is my 3nov version again 2010-11-15T15:32:51 anyway nice work though your bot became awesome quite some time ago 2010-11-15T15:32:58 netjes :) 2010-11-15T15:33:10 eh, it's ok :p 2010-11-15T15:33:51 ^o) 2010-11-15T15:34:02 never seen you in here though Mistmanov 2010-11-15T15:34:22 nope, I knew there was an IRC channel somewhere, but just couldn't find the link 2010-11-15T15:34:28 I see 2010-11-15T15:34:35 woon je in a'dam? 2010-11-15T15:34:41 nope 2010-11-15T15:34:46 sunny hoofddorp 2010-11-15T15:34:53 lol 2010-11-15T15:35:20 i see 2010-11-15T15:35:35 it's annoying that i'm the only serious contender from the VU 2010-11-15T15:35:51 why? 2010-11-15T15:35:56 just.. eh 2010-11-15T15:36:07 would've been nice to have some adversaries that I actually knew 2010-11-15T15:36:12 *** McLeopold has quit IRC (Ping timeout: 276 seconds) 2010-11-15T15:36:28 friend of mine encouraged me to partake in the contest, and then stopped working on his bot like the day after I started 2010-11-15T15:36:33 *** Accoun has joined #aichallenge 2010-11-15T15:36:43 ah ye that sux, I got the same here 2010-11-15T15:37:36 do you study AI at the VU? 2010-11-15T15:37:41 nope 2010-11-15T15:37:59 I figured as much, heard its very non-technical there 2010-11-15T15:38:24 I've had like 1 programming course 2010-11-15T15:39:03 you? 2010-11-15T15:39:14 mmm well you should take more :P 2010-11-15T15:39:23 seem to bed oing well 2010-11-15T15:39:29 I study AI at the UvA 2010-11-15T15:39:39 Ah cool =) 2010-11-15T15:39:43 mistmanov your bot and mine seem to go up against each other quite often 2010-11-15T15:40:13 could be :p 2010-11-15T15:40:51 so what do you study then? 2010-11-15T15:40:55 I don't often look at the official rankings.. outdated version of my bots, and people constantly being thrown to the bottom of the ranking 2010-11-15T15:41:00 (although they fixed that?) 2010-11-15T15:41:13 yes they fixed it 2010-11-15T15:41:14 thrown to the bottom? 2010-11-15T15:41:18 beacuse of new uploads 2010-11-15T15:41:20 Last year in my master Molecular Biophysics 2010-11-15T15:41:22 a with new submissions 2010-11-15T15:41:22 earlier you lost your rank 2010-11-15T15:41:32 now it matches you against similar ranked people even after an upload 2010-11-15T15:41:33 wut, thats quite different 2010-11-15T15:41:36 cool 2010-11-15T15:41:38 I should be writting a report on my internship right now.. 2010-11-15T15:41:43 hahahah 2010-11-15T15:41:47 I'm in the exact same situation 2010-11-15T15:41:58 but then... at my internship they introduced me to this 2010-11-15T15:42:06 bastards 2010-11-15T15:42:12 ah, so that's at least kinda relevant 2010-11-15T15:42:26 ye I guess 2010-11-15T15:42:40 well some of them are participating 2010-11-15T15:42:45 so had to beat them =) 2010-11-15T15:43:14 or die trying or sth 2010-11-15T15:43:21 uhuh ;_ 2010-11-15T15:43:44 yeah the UvA at least has some people uploading bots with some regularity.. 2010-11-15T15:44:08 true, but none in the top 100 -,- 2010-11-15T15:44:20 and i know some study ai 2010-11-15T15:44:27 so they should be :P 2010-11-15T15:44:34 true true =) 2010-11-15T15:44:52 the amount of non-affiliated people in the top rankings surprised me 2010-11-15T15:45:08 well it only became changable after some time 2010-11-15T15:45:15 and most prolly don't care what it says 2010-11-15T15:45:41 *** narnach has joined #aichallenge 2010-11-15T15:46:16 *** wvdschel has quit IRC (Quit: Leaving) 2010-11-15T15:46:18 so ur treating planets as molecules then :P 2010-11-15T15:46:28 nah :p 2010-11-15T15:46:37 I'm not doing anything special at all 2010-11-15T15:46:47 (you could read my excellent threads in the strategy forum ;)) 2010-11-15T15:47:06 oh, haven't done that 2010-11-15T15:47:09 but primarily just calculating "how many ships do I have to send to be sure of victory" 2010-11-15T15:47:19 not a very good approach 2010-11-15T15:47:27 but surprisingly effective 2010-11-15T15:47:32 yeah same here, all the fancy stuff I tried didn't work well or was too much work 2010-11-15T15:47:44 to bring to perfection 2010-11-15T15:48:03 I really expected more bot's like bocksimacko's, that attack in patterns that a human just can't understand 2010-11-15T15:48:11 I honestly have no idea what his bot does 2010-11-15T15:48:34 generates actions and evaluates the gamestate after applying them 2010-11-15T15:48:46 the "max" part in the minimax algorithm 2010-11-15T15:49:02 yeah, so I had just the "programming for newbs" course right ;) 2010-11-15T15:49:14 The problem is (in my experience) getting a proper evaluation functino 2010-11-15T15:49:21 well you can look it up on the wiki 2010-11-15T15:49:24 its not that difficult 2010-11-15T15:50:18 the intro on wikipedia makes it sound easy enough.. 2010-11-15T15:50:21 but ye he created a true monster 2010-11-15T15:50:46 yep 2010-11-15T15:51:04 its the basic AI stuff from year one in the bachelor 2010-11-15T15:51:28 *** amriedle has joined #aichallenge 2010-11-15T15:51:34 however, tuning and tweaking a proper eval function takes a lot of time and effort 2010-11-15T15:51:41 or a very smart person :P 2010-11-15T15:51:47 =) 2010-11-15T15:51:56 effort that would be wasted when the map pool changes btw? ;) 2010-11-15T15:52:13 no not at all 2010-11-15T15:52:13 * Mistmanov opens can of worms 2010-11-15T15:52:32 average growth * average ship count seems to work surprisingly well... 2010-11-15T15:52:52 lavalamp_: for what? 2010-11-15T15:53:00 eval function 2010-11-15T15:53:34 well it highly depends on your actionset I think 2010-11-15T15:53:56 I can't imagine it working well without taking topological factors into account 2010-11-15T15:55:48 oh, I just played you error :p 2010-11-15T15:56:03 crap :P 2010-11-15T15:56:35 another brilliant victory for the VU, the best university in amsterdam ;) 2010-11-15T15:56:40 lol got sniped twice 2010-11-15T15:56:53 *** Utkarsh has quit IRC (Ping timeout: 255 seconds) 2010-11-15T15:56:54 hahahha, ok that's it :P 2010-11-15T15:58:14 *** sigh has quit IRC (Quit: Leaving...) 2010-11-15T15:58:58 vu sux :P 2010-11-15T15:59:06 nothing free about it ^_^ 2010-11-15T15:59:40 -_- 2010-11-15T16:02:17 hehe brb 2010-11-15T16:02:31 *** Utkarsh has joined #aichallenge 2010-11-15T16:05:28 *** davidjliu has joined #aichallenge 2010-11-15T16:07:41 *** irchs has joined #aichallenge 2010-11-15T16:10:31 damn when viewing official games for a while the tcp viewer is slooooooooooooooooooooooooow 2010-11-15T16:10:40 Error323: Yeah I intend to add some. I already have some for my goal selection. 2010-11-15T16:13:46 woot, active bots list: http://zeroviz.us:8080/ 2010-11-15T16:15:28 now for some font changes... (to comic sans) 2010-11-15T16:16:29 NNNOOOOOO 2010-11-15T16:16:30 ;_; 2010-11-15T16:16:38 *** delt0r___ has joined #aichallenge 2010-11-15T16:16:43 oh, the crying emoticon 2010-11-15T16:16:49 I needed that a month ago. 2010-11-15T16:17:25 Top 10 players: bix0r4ever(3791), bocsimacko(3779), protocolocon(3617), _iouri_(3560), dmj111(3559), Mangara(3521), george(3486), shangas(3485), Xrillo(3475), davidjliu(3457) 2010-11-15T16:18:02 *** delt0r has quit IRC (Ping timeout: 255 seconds) 2010-11-15T16:18:23 I know the thingie at the top says "no stupid questions", but uhm.. 2010-11-15T16:18:24 ./tcp zeroviz.us:995 mistmanov -p notmypassword ./MyBot 2010-11-15T16:18:36 returns "no such file or directory" 2010-11-15T16:18:50 McLeopold_, you rock 2010-11-15T16:18:53 no colon? 2010-11-15T16:18:53 um... 2010-11-15T16:19:04 try the ip address 2010-11-15T16:19:18 ip address port number 2010-11-15T16:19:27 ah it was the colon 2010-11-15T16:19:31 thanks =) 2010-11-15T16:25:20 anyone have something that can turn a map file into a png? 2010-11-15T16:25:24 *** Palmik has quit IRC (Remote host closed the connection) 2010-11-15T16:28:21 McLeopold_: in the bit bucket repo, I just dropped some files that will display map files using javascript. not sure if I put everything in there I need.... I have them in another repo at home. 2010-11-15T16:28:51 I'll take a look 2010-11-15T16:29:08 its not pretty... but is how I have been debugging at home. 2010-11-15T16:29:23 create_map_canvas.py and map.js 2010-11-15T16:29:53 I'm thinking of thumbnails for the by_Maps.html 2010-11-15T16:30:01 So you can browse by map type 2010-11-15T16:30:02 sure... don't have that. 2010-11-15T16:30:59 McLeopold_> use javascript 2010-11-15T16:31:24 McLeopold_> png alot of bytes 2010-11-15T16:31:30 ow is that a new server? 2010-11-15T16:31:32 yeah, I could 2010-11-15T16:31:51 I could probably rip out stuff from the visualizer 2010-11-15T16:32:35 who's server is this? 2010-11-15T16:32:38 oh nvm 2010-11-15T16:32:43 i should read up first -.- 2010-11-15T16:35:06 *** p4p4 has quit IRC (Quit: ChatZilla 0.9.84 [SeaMonkey 2.0a3/20090223135443]) 2010-11-15T16:35:56 McLeopold_: what is the actual address? 2010-11-15T16:36:07 zeroviz.us:8080 2010-11-15T16:36:08 for tcp connections 2010-11-15T16:36:14 995 2010-11-15T16:36:22 i tried that... mm 2010-11-15T16:37:05 75.121.152.123 2010-11-15T16:37:07 ./tcp zeroviz.us 995 E323-17.1 -p xxx ./E323-17.1 2010-11-15T16:38:53 76.121.152.123 2010-11-15T16:39:14 *** Mathnerd314 has joined #aichallenge 2010-11-15T16:39:25 *** iamluck has joined #aichallenge 2010-11-15T16:39:39 any luck? 2010-11-15T16:40:10 yes with 76.121.152.123 2010-11-15T16:40:14 not 75 ;) 2010-11-15T16:40:15 cool 2010-11-15T16:40:19 oh 2010-11-15T16:40:22 whoops 2010-11-15T16:40:31 @tcp 2010-11-15T16:40:32 McLeopold_: tcp = (#1) http://www.benzedrine.cx/planetwars/, or (#2) http://72.44.46.68/, or (#3) http://zeroviz.us:8080 2010-11-15T16:41:10 contestbot learn tcp as 76.121.152.123:995 (zeroviz.us) 2010-11-15T16:41:10 McLeopold_: I think that worked... 2010-11-15T16:41:13 @tcp 2010-11-15T16:41:14 McLeopold_: tcp = (#1) http://www.benzedrine.cx/planetwars/, or (#2) http://72.44.46.68/, or (#3) http://zeroviz.us:8080, or (#4) 76.121.152.123:995 (zeroviz.us) 2010-11-15T16:41:41 :) 2010-11-15T16:42:01 *** McLeopold_ has quit IRC (Quit: Leaving) 2010-11-15T16:45:47 http://zeroviz.us down 2010-11-15T16:46:05 :/ 2010-11-15T16:46:38 try again Accoun 2010-11-15T16:58:14 *** Mistmanov has quit IRC (Quit: Page closed) 2010-11-15T17:02:56 wtf double startpositions? O_O 2010-11-15T17:03:06 Yes on TCP server 2010-11-15T17:03:13 I've already seen one 2010-11-15T17:03:23 ye me to just now 2010-11-15T17:04:03 But I think that in the official server, there won't be any map with double startpositions 2010-11-15T17:04:09 http://zeroviz.us:8080/ 2010-11-15T17:04:10 third line 2010-11-15T17:04:33 ye well i read it 2010-11-15T17:04:44 but just didn't expect that :) 2010-11-15T17:05:24 *** sigh has joined #aichallenge 2010-11-15T17:05:24 A friend of mine has done a big generator map with up to 5 map for each player and more than hundred planets 2010-11-15T17:05:43 But I don't think that my bot will perform well on this kind of maps 2010-11-15T17:06:32 *** amriedle has quit IRC (Quit: leaving) 2010-11-15T17:10:17 helpful tip of the day: don't accidentally nano a binary file... 2010-11-15T17:11:59 *** nullkuhl has quit IRC (Read error: Connection reset by peer) 2010-11-15T17:12:58 *** Accoun has quit IRC (Ping timeout: 245 seconds) 2010-11-15T17:17:25 Top 10 players: bocsimacko(3788), bix0r4ever(3722), protocolocon(3568), dmj111(3561), _iouri_(3547), Mangara(3501), shangas(3488), george(3486), asavis(3475), Xrillo(3470) 2010-11-15T17:22:40 oh man! Someone doesn't like having their cheese moved. Well AI is about handling moved cheese. And seriously, it hasn't been moved much at all. 2010-11-15T17:22:47 epic.. 2010-11-15T17:23:01 *** Accoun has joined #aichallenge 2010-11-15T17:27:46 amstan: ehhh? 2010-11-15T17:27:56 quote from the forums 2010-11-15T17:28:19 ah about the new maps i presume, LOL 2010-11-15T17:29:09 ya 2010-11-15T17:29:22 Dude! don't move the cheese! 2010-11-15T17:33:34 *** choas has quit IRC (Ping timeout: 265 seconds) 2010-11-15T17:34:15 *** Accoun has quit IRC (Ping timeout: 255 seconds) 2010-11-15T17:34:27 *** choas has joined #aichallenge 2010-11-15T17:37:45 *** Bobng has quit IRC (Quit: Leaving) 2010-11-15T17:38:34 *** Bobng has joined #aichallenge 2010-11-15T17:40:05 move the cheese more onto the trigger plate 2010-11-15T17:40:31 so that it's easier to trap the mice 2010-11-15T17:41:26 definitely... point a laser at the cheese 2010-11-15T17:41:38 boobietrapped planets 2010-11-15T17:41:42 kaboom 2010-11-15T17:42:40 wow, it looks like my old code handles the new maps far better than my new code 2010-11-15T17:43:27 *** lavalamp_ has quit IRC (Quit: Page closed) 2010-11-15T17:43:47 *** Accoun has joined #aichallenge 2010-11-15T17:45:07 *** choas has quit IRC (Quit: leaving) 2010-11-15T17:50:04 woah. 16 game winning streak 2010-11-15T17:50:05 how exciting 2010-11-15T17:50:15 except apparently I like losing to people outside the top 100 2010-11-15T17:56:29 Just to know, how much line is your bot around? 2010-11-15T17:56:55 how much line? 2010-11-15T17:57:30 yes, merging all the files, how much lines of code does that make 2010-11-15T17:57:38 Oh I see 2010-11-15T17:57:56 (my english is still bad, sorry) 2010-11-15T17:58:16 Well, I guess it doesn't help you much unless you use the same language as me 2010-11-15T17:58:31 is there an assembly pack? performance ftw! 2010-11-15T17:58:40 burny, you can make one :) 2010-11-15T17:58:59 I'm using Python and I have about 800 lines of code 2010-11-15T17:59:04 I'm using python too 2010-11-15T17:59:28 Good man 2010-11-15T17:59:35 But then, my bot isn't really all that good 2010-11-15T17:59:48 I use c++ and I have about 3 times that. Though only about 1000 lines actual bot code, the other are support classes 2010-11-15T17:59:53 I have something about 1400 lines of code, but with a lot of comments and all 2010-11-15T18:00:14 by the way what does it matter? 2010-11-15T18:00:31 It doesn't - I think medrimonia was just curious 2010-11-15T18:00:55 Just curious 2010-11-15T18:01:14 k 2010-11-15T18:01:38 And wanting to know if I'm not searching too much complicated 2010-11-15T18:01:59 actually I think I have more like 5000 lines of code, but I write { and } on their own lines 2010-11-15T18:02:04 (usually) 2010-11-15T18:02:12 Yeah 2010-11-15T18:02:17 mine's around 700 for what I've written, not including the unofficial starter pack, but that's including comments and stuff 2010-11-15T18:02:22 there's probably a top10 bot that can be done in around 100 lines 2010-11-15T18:02:42 well, I can make my bot 1 line. 2010-11-15T18:02:49 It'll vary hugely depending on the language and coding style anyway 2010-11-15T18:02:50 and get top 10? 2010-11-15T18:02:52 Migi32: do it. 2010-11-15T18:02:59 nah, not get top 10 :P 2010-11-15T18:03:55 Who wrote the starter packs? 2010-11-15T18:04:10 The Python one is coded in a rather non-python fashion 2010-11-15T18:04:11 d0ugal, that would require removing all #include's and copy-pasting and all, I won't get into that. 2010-11-15T18:04:19 ha 2010-11-15T18:04:48 *** McLeopold has joined #aichallenge 2010-11-15T18:04:55 Suppose I could do my bot as one line of code with an eval and loads of \n's 2010-11-15T18:05:38 debuging would be uber fun 2010-11-15T18:05:46 error on line 1 :P 2010-11-15T18:06:17 :D 2010-11-15T18:06:24 i'm about 1000 lines, but not sure how much of it is even used any more :/ 2010-11-15T18:06:41 who made zeroviz.us? 2010-11-15T18:06:45 heh 2010-11-15T18:07:07 Migi32: McLeopold 2010-11-15T18:07:19 nice! :) 2010-11-15T18:09:25 McLeopold, about the maps with multiple starts, how about making it like this: port 8080 = "standard" maps, rated. port 8081 = special maps, unrated 2010-11-15T18:10:05 I thought about that, but then I thought maybe enough people wouldn't connect 2010-11-15T18:10:48 What I'd really like is for a bot to be able to request which map types they are willing to play 2010-11-15T18:11:20 make an optional #define 2010-11-15T18:12:28 McLeopold, well you could do that with comments, but there you will most likely not have enough people 2010-11-15T18:12:48 yeah, I'll think about the multiple port idea 2010-11-15T18:13:12 would it be possible to have port 8081 play against both other people connected to 8081 *and* people connected to 8080 (on regular maps then)? 2010-11-15T18:13:23 that would work 2010-11-15T18:13:51 ok, it was just an idea, do with it what you want :) but I have to go now 2010-11-15T18:13:52 bye :) 2010-11-15T18:13:56 *** Migi32 has quit IRC (Quit: www.ai-contest.com) 2010-11-15T18:14:50 or port 8082 is one type, 8083 is the other.. and 8081 can switchbetween 82+83 2010-11-15T18:17:25 Top 10 players: bocsimacko(3785), bix0r4ever(3668), dmj111(3559), _iouri_(3556), protocolocon(3540), shangas(3494), george(3492), davidjliu(3468), jambachili(3466), Xrillo(3466) 2010-11-15T18:19:11 I would want the results from all ports to be included together. The current server.cpp would play nice with multiple instances. 2010-11-15T18:19:28 So, it probably won't happen until after the official contest. 2010-11-15T18:19:56 If there is a consensous to only use official maps, then I will remove the multiple start maps. 2010-11-15T18:20:14 janzert: how many maps are running on the official server? 2010-11-15T18:20:31 *** JamesMG has quit IRC (Quit: Leaving.) 2010-11-15T18:21:10 *** irchs has quit IRC (Quit: irchs) 2010-11-15T18:22:24 *** Bobng has quit IRC (Ping timeout: 255 seconds) 2010-11-15T18:43:10 *** McLeopold has quit IRC (Quit: Leaving) 2010-11-15T18:44:51 *** fawek has quit IRC (Ping timeout: 276 seconds) 2010-11-15T18:57:20 *** Rubicon-|-Cross has quit IRC (Quit: Leaving) 2010-11-15T18:57:54 *** jaspervdj has quit IRC (Quit: NEVER GONNA GIVE YOU UP NEVER GONNA LET YOU DOWN) 2010-11-15T19:17:25 Top 10 players: bocsimacko(3775), bix0r4ever(3647), dmj111(3565), protocolocon(3564), _iouri_(3541), shangas(3501), asavis(3499), george(3493), _Astek_(3490), jambachili(3486) 2010-11-15T19:43:28 *** superflit has quit IRC (Quit: superflit) 2010-11-15T19:43:47 *** superflit has joined #aichallenge 2010-11-15T19:47:17 *** phreeza has quit IRC (Quit: Computer has gone to sleep) 2010-11-15T19:48:21 *** shades has quit IRC (Ping timeout: 255 seconds) 2010-11-15T20:05:32 *** McLeopold has joined #aichallenge 2010-11-15T20:16:04 *** McLeopold has quit IRC (Quit: Leaving) 2010-11-15T20:16:10 *** mceier has quit IRC (Quit: leaving) 2010-11-15T20:17:31 Top 10 players: bocsimacko(3792), bix0r4ever(3624), dmj111(3576), protocolocon(3573), _Astek_(3549), _iouri_(3545), shangas(3510), george(3490), jambachili(3485), Xrillo(3484) 2010-11-15T20:52:38 *** nullkuhl has joined #aichallenge 2010-11-15T20:53:49 *** McLeopold has joined #aichallenge 2010-11-15T20:54:11 antimatroid: you there? 2010-11-15T20:55:11 yes 2010-11-15T20:55:13 what's up? 2010-11-15T20:55:25 there are 30 maps on the main server being used atm btw 2010-11-15T20:57:41 McLeopold: ? :P 2010-11-15T21:10:56 *** rebelxt_ has joined #aichallenge 2010-11-15T21:16:31 What does "error: invalid use of incomplete type..." mean? 2010-11-15T21:17:30 nvm 2010-11-15T21:17:31 Top 10 players: bocsimacko(3777), bix0r4ever(3636), protocolocon(3589), dmj111(3577), _iouri_(3560), shangas(3509), george(3486), Xrillo(3478), Raschi(3457), jambachili(3456) 2010-11-15T21:17:37 Should have googled first :/ 2010-11-15T21:33:33 antimatroid: is there only 30 maps on the official right now? 2010-11-15T21:34:21 I think that's what janzert said 2010-11-15T21:34:40 are you going to generate more? 2010-11-15T21:35:06 i expect he will, i don't do any of that stuff 2010-11-15T21:35:09 i'm still competing :) 2010-11-15T21:35:22 i have 2 ideas for more maps on your server 2010-11-15T21:35:28 what are they? 2010-11-15T21:36:04 1. start with no neutrals, but all symmetric 2010-11-15T21:36:04 2. all neutrals closest to you are non-snipable after the first move if take just the one, and centre planets start with ships <= their growth level 2010-11-15T21:36:58 hmm 2010-11-15T21:36:58 i'll knock generators up for those later if you'd be keen to add some? 2010-11-15T21:37:18 That would be interesting. 2010-11-15T21:37:33 But, I'm thinking of only using maps that would be used in the finals. 2010-11-15T21:37:38 :( 2010-11-15T21:37:48 That's why I'm wondering if there are more "official" maps. 2010-11-15T21:38:05 I'd like to focus on tuning for the contest. 2010-11-15T21:38:08 i wouldn't put up the same set the official server is using 2010-11-15T21:38:11 antimatroid: set up your own server if you want to continue experimenting with map generators 2010-11-15T21:38:12 just use the generator 2010-11-15T21:38:32 I've already put the official maps on. 2010-11-15T21:38:35 I think 2010-11-15T21:39:36 McLeopold> http://code.google.com/p/ai-contest/source/browse/trunk/planet_wars/backend/map_generator.py# 2010-11-15T21:39:56 thx 2010-11-15T21:40:25 McLeopold: great, thanks. i like your use of jquery 2010-11-15T21:41:16 :) 2010-11-15T21:41:29 The site was assembled, not written. 2010-11-15T21:41:49 antimatroid: you wrote that generator, right? 2010-11-15T21:41:58 yes, janzert made a couple of changes 2010-11-15T21:42:07 did you also re-use dhartmei's server code like i did? 2010-11-15T21:42:36 yeah, but I had to fix it for non-bsd 2010-11-15T21:43:11 *** bhasker_ has quit IRC (Ping timeout: 264 seconds) 2010-11-15T21:43:12 there are 2 changes you could make now, make maxRadius 17 and change line 184 to num_ships = min(maxShips, random.randint(minShips, 5*distance(p1, p2)-1)) 2010-11-15T21:43:40 just noticed "Active bots" table - sweet! 2010-11-15T21:44:04 k 2010-11-15T21:44:15 does it randomly pick line or radial? 2010-11-15T21:44:22 i wonder how bots will go with 30 planets and 34 max distance 2010-11-15T21:44:26 i don't think that's been done yet 2010-11-15T21:44:32 yeah 2010-11-15T21:44:38 should be ~50% split 2010-11-15T21:45:15 a set of 100 new ones would probably be nice to start with 2010-11-15T21:45:23 then add more once people have sort of seen a lot of them 2010-11-15T21:46:00 I may setup up a second port for fun maps, so go ahead and do that generator 2010-11-15T21:46:21 i will do it later today, i want to play them anyway 2010-11-15T21:46:25 even if others don't 2010-11-15T21:47:36 what's the change to line 184 for? 2010-11-15T21:47:50 so a neutral never starts with ships over 100 2010-11-15T21:47:58 it's more for aesthetic purposes than anything 2010-11-15T21:48:28 that's a complicated generator 2010-11-15T21:48:48 I'll put mine on bitbucket later so you can see how easy it should be :P 2010-11-15T21:48:57 McLeopold: separate port for fun maps is a good idea. can you keep ratings pages separate as well then? 2010-11-15T21:49:03 Yes 2010-11-15T21:49:40 awesome 2010-11-15T21:50:28 yeah, I had never used python before and made if before i even knew wtf radial and line symmetry are called etc. 2010-11-15T21:51:16 i hope it's not just me and McLeopold playing them aha 2010-11-15T21:59:49 *** mega1 has joined #aichallenge 2010-11-15T22:10:55 *** amstan has quit IRC (Remote host closed the connection) 2010-11-15T22:12:23 *** amstan has joined #aichallenge 2010-11-15T22:12:23 *** ChanServ sets mode: +o amstan 2010-11-15T22:14:15 amstan 2010-11-15T22:14:21 McLeopold 2010-11-15T22:14:32 I've got a very serious question 2010-11-15T22:14:54 ok? 2010-11-15T22:14:55 Is it a possibility for the dealine to be extended? 2010-11-15T22:15:16 a lot more people would complain about that than the mapgen 2010-11-15T22:15:30 So that's a no? 2010-11-15T22:16:01 I think the new maps should only be used if there is an extension. 2010-11-15T22:16:13 Otherwise, we should stick to the original generator. 2010-11-15T22:16:27 why? 2010-11-15T22:16:31 like 2010-11-15T22:16:33 see that topic 2010-11-15T22:16:35 see the poll 2010-11-15T22:16:53 I like the new maps, but even I had unconsiously optimized for the older style. 2010-11-15T22:17:27 Top 10 players: bocsimacko(3771), bix0r4ever(3633), dmj111(3576), protocolocon(3564), _iouri_(3559), davidjliu(3515), shangas(3505), george(3487), Xrillo(3477), jambachili(3458) 2010-11-15T22:17:32 The game is now more different than it was. 2010-11-15T22:18:05 you did that under assumptions 2010-11-15T22:18:10 I would prefer and extension. 2010-11-15T22:18:19 No, I didn't. 2010-11-15T22:18:30 I made a lot of crazy test maps. 2010-11-15T22:19:05 amstan: i think in future you should state that a bran new generator will be used in the final tournament, that only organisers will know the details of 2010-11-15T22:19:16 Anyway, consider a second vote now that people have experienced the new maps. 2010-11-15T22:19:35 so like.. bother jeff with this 2010-11-15T22:19:44 antimatroid: I dissagree, openess is more fair. 2010-11-15T22:19:55 open the generator at the end 2010-11-15T22:20:03 *** superflit has joined #aichallenge 2010-11-15T22:20:06 or release a brand new one with a week left, so errors can be found 2010-11-15T22:20:21 but i don't like having to gear my bot for a specific generator, that's boring 2010-11-15T22:20:59 I'm just giving my opinion. 2010-11-15T22:21:07 as am i :) 2010-11-15T22:21:09 I thought you should know. 2010-11-15T22:22:31 i haven't made a single choice off results of games, my bot is entirely written to implement a given strategy, everything is a matter of getting it to do what i intended 2010-11-15T22:22:42 i would expect that's how such a contest should really be done 2010-11-15T22:22:50 well, that's my idea of a perfect ai contest anyway :) 2010-11-15T22:23:12 McLeopold: you probably saw it on the forums that i had the same reaction but had no luck reversing the change. so i had to change my bot significantly since my eval function sucked for line symmetric maps 2010-11-15T22:23:47 Yeah, I suspect a lot of people will have to rethink heuristics. 2010-11-15T22:24:00 That's why I would prefer more time, or original maps. 2010-11-15T22:24:18 my vote is with more time over original maps 2010-11-15T22:24:23 me too! 2010-11-15T22:24:45 but i really don't see why, people were never told to gear their heuristics towards point symmetric maps 2010-11-15T22:24:53 maybe no example generator should be given in future? 2010-11-15T22:25:03 tell contestants to come up with their own examples entirely 2010-11-15T22:25:11 antimatroid: i think the issue was more of the maps already on the server 2010-11-15T22:25:25 amstan: what do you mean? 2010-11-15T22:25:39 antimatroid: we have to have maps on the servers already 2010-11-15T22:25:41 for the visualizer 2010-11-15T22:26:02 i'm confused 2010-11-15T22:26:11 oh wait 2010-11-15T22:26:21 yeah, but they can be stated as not representative of final maps either 2010-11-15T22:26:30 it's just not said that they are now 2010-11-15T22:27:18 at least there is a clear definition of the limits of the map now, but it was a little late 2010-11-15T22:27:30 better late than never :P 2010-11-15T22:27:50 horseshoes and handgrenades 2010-11-15T22:28:05 i've seen that expression twice today now 2010-11-15T22:28:10 first two times ever 2010-11-15T22:28:11 why? 2010-11-15T22:28:23 are you familiar with it? 2010-11-15T22:28:38 the other one gave more details about it being the only approximaion method or something? 2010-11-15T22:28:40 other than that, no 2010-11-15T22:28:58 McLeopold: i'm actually glad you came to this conclusion too even though it might be too late to reverse the map change decision 2010-11-15T22:29:10 close only counts in horseshoes and handgrenades 2010-11-15T22:29:24 meaning, you should be more exact 2010-11-15T22:29:43 more exact about it being better late than never? 2010-11-15T22:29:46 like, not waiting 80% into a contest to throw a curveball 2010-11-15T22:29:57 it's not a curveball when the map generator was meant to change 2010-11-15T22:30:09 my only issue is that it could have been clearer 2010-11-15T22:30:19 and earlier 2010-11-15T22:30:33 but people can't complain about the timing that much, they could have pushed like i did earlier 2010-11-15T22:30:35 I agree that it is within the rules. But there was no clear definition before. 2010-11-15T22:31:08 Anyway, I know how to deal with the new maps, I just have to do it now. 2010-11-15T22:31:16 I'm thinking of the others... 2010-11-15T22:31:29 do you understand my argument though? 2010-11-15T22:31:46 yeah, but I dissagree 2010-11-15T22:31:47 that i haven't worked under those assumptions wheras some have, it's unfair to then just keep the old maps as well 2010-11-15T22:31:49 why? 2010-11-15T22:32:00 time extension would be nice but i don't think this is very realistic. think of all the top people who want their life back ;-) 2010-11-15T22:32:01 should i have made assumptions and got fucked over if most people didn't? 2010-11-15T22:32:20 with a lack of clear guidelines, people had to make assumptions 2010-11-15T22:32:50 my only implicit assumption is that i couldn't handle planet sizes over like 50 or 60 2010-11-15T22:32:59 i don't tihnk assumptions past that needed to be made 2010-11-15T22:33:13 should I have made more assumptions? I think that's a terrible thing to do 2010-11-15T22:33:21 yeah, but you still made it, because it wasn't stated 2010-11-15T22:33:33 yeah, but i could probably fix it to work on larger maps 2010-11-15T22:33:39 it'd just be less smart 2010-11-15T22:33:50 even that i'm not entirely bound to 2010-11-15T22:33:59 strictly because of the specifications 2010-11-15T22:34:13 i followed the rules, and you're saying i should be punished because a majority didn't 2010-11-15T22:34:22 That's not what I'm saying 2010-11-15T22:34:27 i think majority based decision rules are a crock of shit :P 2010-11-15T22:34:53 what are you saying then? 2010-11-15T22:35:32 Because there was not a clear definition of what a map would be, the organizers should stick to the original generator that people made their assumptions off of. 2010-11-15T22:35:49 what about the people who worked off what the specification actually said? 2010-11-15T22:35:59 or more importatly, didn't say? 2010-11-15T22:35:59 what specification? 2010-11-15T22:36:08 http://ai-contest.com/specification.php 2010-11-15T22:36:24 "Maps have no particular dimensions and are defined completely in terms of the planets and fleets in them." 2010-11-15T22:36:24 "About the Current MapsMost maps on the server and in the starter packs were generated programmatically by a Python script. You may also create your own to use for testing." 2010-11-15T22:36:50 people filled in the lines on the rest 2010-11-15T22:38:14 it also doesn't say the turn limit will be specified, while I use that, i'm ready to throw that out in about 30 seconds if needed 2010-11-15T22:38:21 it sounds like a lot of people DEPEND on it 2010-11-15T22:39:20 I'm not sure if we have an official 200 yet, do we? 2010-11-15T22:39:34 i highly doubt it will change, but it isn't official 2010-11-15T22:39:43 amstan: can you just make that official and shut people up 2010-11-15T22:39:54 :) 2010-11-15T22:40:02 sure.. 200 2010-11-15T22:40:19 while i have had games that I think would have changed results, they're very few and far between 2010-11-15T22:50:02 http://28.media.tumblr.com/tumblr_lbxrvcK4pk1qbylvso1_400.png that's great 2010-11-15T22:50:22 *** pgpaskar1 is now known as pgpaskar 2010-11-15T22:51:33 antimatroid: yeah, the canonical version of that is to say the diagonal of a square is of length 2 2010-11-15T22:52:15 thing is it's not entirely wrong 2010-11-15T22:52:24 it depends which norm you're using 2010-11-15T22:52:34 yes it is, derivative does not converge 2010-11-15T22:52:48 people assume you're using the euclidean norm 2010-11-15T22:52:56 or metric or whatever 2010-11-15T22:53:02 ok, run away into your fancy spaces then :P 2010-11-15T22:53:09 i will :) 2010-11-15T22:55:16 *** amriedle has joined #aichallenge 2010-11-15T22:57:21 anyone is hiden popular TCP server with new map set ? 2010-11-15T22:58:56 sorry, what? 2010-11-15T22:59:12 oh, yeah, I don't think new maps are on there 2010-11-15T22:59:25 although mcleopolds server is slowly getting recognised 2010-11-15T22:59:38 okay, not at the moment it's not 2010-11-15T23:00:44 so when is the fixed new generator expected? 2010-11-15T23:02:53 *** lavalamp_ has joined #aichallenge 2010-11-15T23:03:06 Accoun, you've got a bug: http://zeroviz.us:8080/canvas?game_id=8975 2010-11-15T23:03:44 I'm working on 100 new maps for my tcp server 2010-11-15T23:04:00 I'm working on a bot worth running on your server :/ 2010-11-15T23:04:11 I'm referring to the map area/widht issue 2010-11-15T23:04:23 the radius 17 2010-11-15T23:04:42 it would be really cool to have the generator finalized 2010-11-15T23:04:45 mega1: Janzert is out of town, so after that i'd say 2010-11-15T23:05:08 antimatroid: is the code updated? 2010-11-15T23:05:19 if so I generate a few myself. 2010-11-15T23:05:27 no point in testing on the old ones 2010-11-15T23:05:34 hang on, i'll update one and post it on code pad 2010-11-15T23:05:41 and tell you what i changed 2010-11-15T23:05:51 *** TheMacca has joined #aichallenge 2010-11-15T23:06:01 *** bhasker has joined #aichallenge 2010-11-15T23:06:20 better yet, you may also want to post a sample, say 200 2010-11-15T23:08:21 gah, that official one wont even run on my computer now :| 2010-11-15T23:10:12 mega1: if the official one works for you, mapRadius should really be 17 and line 184 should be 2010-11-15T23:10:12 num_ships = max(maxShips, random.randint(minShips, 5*distance(p1, p2)-1)) 2010-11-15T23:10:12 to ensure the maxShips on a neutral condition is met 2010-11-15T23:10:56 although some people seem to think there's a difference between having the map generated inside a square and circle with different areas, your thoughts? I thought just the maxDistance was important 2010-11-15T23:12:44 I haven't thought about the details at all. 2010-11-15T23:12:57 I'll gen some maps 2010-11-15T23:13:12 * lavalamp_ makes anti's changes 2010-11-15T23:13:22 okay, 100 new maps on my server, with the generator changes 2010-11-15T23:13:33 :) 2010-11-15T23:13:44 I hope we don't crash. :) 2010-11-15T23:14:43 btw, There are high winds at my house. This server may not be up all night. 2010-11-15T23:14:52 test now! 2010-11-15T23:15:17 ah ok 2010-11-15T23:15:30 first new map game: http://zeroviz.us:8080/canvas?game_id=9141 2010-11-15T23:16:03 hmm, 4 100 ships planets... 2010-11-15T23:16:14 antimatroid: you might need to double check your numbers 2010-11-15T23:16:23 what do you mean? 2010-11-15T23:16:29 it seems high 2010-11-15T23:16:42 http://ai-contest.com/forum/viewtopic.php?f=19&t=1094&p=7353#p7353 2010-11-15T23:16:43 too high to ever be taken? 2010-11-15T23:16:51 200 maps there 2010-11-15T23:17:15 nope, apparently not :P 2010-11-15T23:17:24 it shouldn't happen that often 2010-11-15T23:17:27 Top 10 players: bocsimacko(3754), bix0r4ever(3563), davidjliu(3562), dmj111(3558), _iouri_(3552), protocolocon(3501), SUA.khb(3497), shangas(3484), george(3471), Xrillo(3454) 2010-11-15T23:17:33 yay my dumb bot :/ 2010-11-15T23:18:13 mega1 I think I actually made 200 this time and not 199 ;) 2010-11-15T23:19:12 lavalamp_: thank you, I should be in bed, but tomorrow I'll test them 2010-11-15T23:19:21 is it with the changes antimatroid outlined? 2010-11-15T23:19:27 Yeah, those two changes. 2010-11-15T23:19:28 yes 2010-11-15T23:20:12 here's the exact script I used: http://pastebin.com/zBafpvuM 2010-11-15T23:20:16 i'd be curious of any others thoughts on whether it matters that the space in which planets can be placed is now larger but the max distance is preserved? 2010-11-15T23:20:31 density matters 2010-11-15T23:20:38 you basically having a circle encapsulating the old square 2010-11-15T23:20:43 density should be near even 2010-11-15T23:20:57 radius is calculated as r = sqrt(rand(0,1))*maxRadius 2010-11-15T23:21:08 but not the same as in the old generator, right/ 2010-11-15T23:21:08 ? 2010-11-15T23:21:09 distribution sorry 2010-11-15T23:21:19 that was generated with cartesian coordinates 2010-11-15T23:21:21 *** davidjliu has quit IRC (Quit: Page closed) 2010-11-15T23:21:32 square might leave some planets in the extreme corners, more outliers. Important? not sure 2010-11-15T23:21:54 if it is considered important, that could be fixed, my generator is pretty horrible :P 2010-11-15T23:21:58 but it works 2010-11-15T23:23:15 will the visualizer support the larger circle? 2010-11-15T23:23:34 no idea how it works 2010-11-15T23:23:55 if it's just giving a square of dimensions just above 17x17, then no 2010-11-15T23:24:33 that wouldn't be hard to fix though, just scan the planets to generate max/min x/y values 2010-11-15T23:24:37 games like this give me hope: http://zeroviz.us:8080/canvas?game_id=9143 2010-11-15T23:25:13 ouch: http://zeroviz.us:8080/canvas?game_id=9154 2010-11-15T23:25:13 *** bhasker has quit IRC (Read error: Connection reset by peer) 2010-11-15T23:25:26 hehe 2010-11-15T23:25:35 well i still think the new maps are awesome 2010-11-15T23:25:43 *** bhasker has joined #aichallenge 2010-11-15T23:28:34 McLeopold I think you found a bug in the code that garantees there's a neutral you can capture... it should verify that the neutral is not between you and the other guy... 2010-11-15T23:29:14 tell antimatroid 2010-11-15T23:29:31 lavalamp_: does it matter? 2010-11-15T23:29:38 then you just have their one you can get 2010-11-15T23:29:45 oh wait, yeah okay 2010-11-15T23:29:55 it means neither player can take it without getting sniped 2010-11-15T23:30:03 it was in the new spec, but only specified against ragebot 2010-11-15T23:30:07 nothing is said about not being sniped 2010-11-15T23:30:12 so that's a draw map 2010-11-15T23:30:16 lol, good point 2010-11-15T23:30:19 although i think that could be added 2010-11-15T23:30:21 there are a few that are drawish 2010-11-15T23:30:23 I don't care if there's draw maps 2010-11-15T23:30:39 they weed out the stupid 2010-11-15T23:30:48 there's no really bad draw maps anymore at least 2010-11-15T23:30:59 i don't think 2010-11-15T23:31:13 I thought about the maps in the final tournament 2010-11-15T23:32:24 my feeling is that how result of a bot on a particular map depend depends a lot more on the map than the map+opponent 2010-11-15T23:32:43 yes, map type is a *big* factor 2010-11-15T23:32:44 *how well a bot plays on a 2010-11-15T23:32:52 hasn't that always been the case? 2010-11-15T23:32:56 yes 2010-11-15T23:32:59 yes 2010-11-15T23:33:36 this makes me think that for a round robin of a hundred players we generate 200 maps and each player plays each map once 2010-11-15T23:33:51 it's pretty much impossible to remove that element from a contest, I think that comes back to the tradeoff from having too few and too many maps 2010-11-15T23:34:04 mega1: i would like something like that too, although the numbers might be a bit high to be feasible there 2010-11-15T23:34:11 i wanted something like that for tron too 2010-11-15T23:35:02 the biggest problem is your final top x are determined by games only with people inside that range, but i think it fixes more problems than that personally 2010-11-15T23:35:41 you can replace a huge round robin with a series of small ones 2010-11-15T23:35:56 i started a discussion on this, did any of you see it? 2010-11-15T23:36:05 the light are flickering... 2010-11-15T23:36:20 how windy does it get tehre? 2010-11-15T23:36:25 i.e., 1-10 play a round robin, 11-20 play one, 21-30, etc. Then the table winners play a final round robin. 2010-11-15T23:36:51 13mph 2010-11-15T23:36:58 actually you should use slaughter parings 2010-11-15T23:36:59 i think having the final tournament run in rounds incrementally chopping off bottom sections would work reasonably well too 2010-11-15T23:37:18 so 1, 11, 21 ... 91 play, and 2, 12, 22... play, etc 2010-11-15T23:37:19 but the bottom bots don't get a very accurate ranking 2010-11-15T23:37:27 other option is mergesort like tournaments 2010-11-15T23:37:48 that way the last round is more likely to have the strongest players 2010-11-15T23:37:49 with the final round being a round robbin for each group 2010-11-15T23:39:01 *** McLeopold has quit IRC (Quit: Leaving) 2010-11-15T23:39:01 i think they need to get a copy of the top 1000 bots at the end of this, and test out a bunch of different ranking methods, and try to work out how each one does somehow 2010-11-15T23:39:30 again, this should have specified at the beginning 2010-11-15T23:39:34 +been 2010-11-15T23:39:53 no arguments here on that :P 2010-11-15T23:40:18 i think i'm in favour of more general rules than most, but everybody agrees that the rules should be clear at the start 2010-11-15T23:40:32 for this reason I suggest sticking to how the tron tournament was done as much as possible 2010-11-15T23:40:43 i don't like the way pairings was done for that 2010-11-15T23:40:45 because that's the only thing contestants had to go by 2010-11-15T23:40:50 how was it? 2010-11-15T23:40:50 the top bots didn't play each other enough 2010-11-15T23:40:57 completely random 2010-11-15T23:41:07 go back and look how often you played the top 10 2010-11-15T23:41:11 I thought it was a round-robin for the top 100 2010-11-15T23:41:16 don't think so 2010-11-15T23:41:59 well, there was no round robbin, not enough games 2010-11-15T23:42:30 http://csclub.uwaterloo.ca/contest/profile_games.php?user_id=2788 2010-11-15T23:42:36 that was my wonderful record till i timed out 2010-11-15T23:43:33 *** bhasker has quit IRC (Read error: Connection reset by peer) 2010-11-15T23:44:14 *** bhasker has joined #aichallenge 2010-11-15T23:44:33 a1k0n played 2-5 exactly once 2010-11-15T23:44:53 together with my vague recollection this is suggestive 2010-11-15T23:45:03 hmm okay 2010-11-15T23:45:10 i'd prefer to have the top bots playing each other quite a lot 2010-11-15T23:45:48 otherwise 1 and 2 might be decided just off the luck of the map they were played on 2010-11-15T23:46:01 that would be bad for someone like you who is obviously the best so far 2010-11-15T23:46:12 and i think it's unideal in general 2010-11-15T23:46:55 with a top x round robin, what if each player was assigned a map, and they played every other bot on that map? 2010-11-15T23:47:16 nah that's terrible ignore that 2010-11-15T23:47:20 that's about the worst possible solution 2010-11-15T23:47:29 :) 2010-11-15T23:47:59 gah finished rewriting the state code with tuples only to realize tuples are immutable 2010-11-15T23:48:00 maybe if they also played only one opponent 2010-11-15T23:48:01 !@#!@#!@# 2010-11-15T23:48:10 ehehe 2010-11-15T23:48:18 have to restructure code to not try and update tuples 2010-11-15T23:48:36 I'm off to bed. 2010-11-15T23:48:45 please beat up b0.64 on zeroviz 2010-11-15T23:48:48 dream up a perfect ranking system :) 2010-11-15T23:49:02 http://72.44.46.68/getplayer?player=bocsimacko-b0.64 2010-11-15T23:49:07 mega1 I'll do my best xP 2010-11-15T23:49:07 looks like as-z03 2010-11-15T23:49:08 did it 2010-11-15T23:49:08 it should be better at line symmetric maps 2010-11-15T23:49:09 already 2010-11-15T23:49:18 on tcp 2010-11-15T23:49:40 that was a point symmetric map anyway 2010-11-15T23:49:40 no sleep for you mega1 2010-11-15T23:49:42 bhasker: I'm more interested in zeroviz 2010-11-15T23:49:45 ah k 2010-11-15T23:51:43 iouri up to second 2010-11-15T23:51:48 *** Utkarsh has quit IRC (Ping timeout: 272 seconds) 2010-11-15T23:55:50 bah, zeroviz's down 2010-11-15T23:56:33 must have got too windy 2010-11-15T23:59:50 *** Utkarsh has joined #aichallenge