2010-09-29T00:03:23 *** Naith has quit IRC (Client Quit) 2010-09-29T00:08:51 woot above 3000 2010-09-29T00:08:56 http://ai-contest.com/profile.php?user_id=5822 2010-09-29T00:09:10 very nice 2010-09-29T00:09:37 and looks like my bot is doing much better on tcp as well 2010-09-29T00:09:41 @36 position now 2010-09-29T00:09:42 bhasker: No! 2010-09-29T00:09:51 ... 2010-09-29T00:11:02 haha, shut down 2010-09-29T00:12:57 _iouri_'s presence doesn't surprise me. his bot was strong last time 2010-09-29T00:13:07 bhasker: get a lot done working from home today? :) 2010-09-29T00:13:46 haha actually my current bot is barely changed:( 2010-09-29T00:13:50 i have to rewrite it 2010-09-29T00:13:59 i just fixed a couple of heuristics 2010-09-29T00:14:03 and a bug i had in the code 2010-09-29T00:14:04 nice, hasker 2010-09-29T00:14:29 * Sylph nods 2010-09-29T00:14:42 Sylph nods a lot :P 2010-09-29T00:14:57 i went through a lot of revisions on my bot with nothing to show for it yet 2010-09-29T00:15:04 *** keynes has joined #aichallenge 2010-09-29T00:15:05 though i believe that my current bot is doing okay on tcp only because a lot of the bots in the top 30 are not playing right now 2010-09-29T00:15:06 but i think i have an idea of what i want to do now at least 2010-09-29T00:15:46 I'm a very very agreeable person, sigh. :D 2010-09-29T00:15:47 i wish i had a more mathematical bent of mind:( 2010-09-29T00:15:50 hrm 2010-09-29T00:15:50 a1k0n: it took you this long to figure out you wanted to win? :D 2010-09-29T00:15:56 nm 2010-09-29T00:16:09 i meant more someone who like to agree and stuff :P 2010-09-29T00:16:12 heh. i was drawing a blank on an actual approach for a while 2010-09-29T00:16:19 mmm 2010-09-29T00:17:01 what rank are oyu now a1kon ? 2010-09-29T00:17:16 omg wednesday already :| 2010-09-29T00:17:17 hrm 2010-09-29T00:17:22 how am i going to get time to code my bot >.< 2010-09-29T00:17:24 when you upload a new bot why doesnt it fight someone around the level of your old bot? 2010-09-29T00:17:32 not sure 2010-09-29T00:17:34 that'll be helpful 2010-09-29T00:17:39 not wasting time :/ 2010-09-29T00:17:44 yea 2010-09-29T00:17:56 Sylph: dunno, it keeps bouncing around like crazy 2010-09-29T00:18:01 ah 2010-09-29T00:18:04 it don't stabilize ? 2010-09-29T00:18:06 i haven't uploaded since the day the contest started 2010-09-29T00:18:06 that's on the list of stuff to be fixed, I think 2010-09-29T00:18:11 ah 2010-09-29T00:18:12 not... really 2010-09-29T00:18:24 mine seems to staabilize at slightly above 300 2010-09-29T00:18:29 (playing bots your own level for new submission) 2010-09-29T00:18:30 and stays around there 2010-09-29T00:18:58 a1k0n is sandbagging, he was in the top 10 earlier 2010-09-29T00:19:10 dude i wish 2010-09-29T00:19:16 :P 2010-09-29T00:19:33 there is no rational explanation for why a1k0n-0.2 is any good 2010-09-29T00:19:41 but that's what's up there 2010-09-29T00:20:13 i just realized this morning that october comes before november 2010-09-29T00:20:22 and thus we have a lot of time left 2010-09-29T00:20:28 :) 2010-09-29T00:20:52 :P 2010-09-29T00:20:58 so much more time 2010-09-29T00:21:52 I need to make something so that my bot can optimise parameters by itself and leave it going for a few weeks 2010-09-29T00:23:28 whats with the invalid input on the server 2010-09-29T00:24:48 sigh: we're planning to do something like that, still not entirely sure how 2010-09-29T00:25:04 as in how to update the parameters, pretty obvious "how" to do it 2010-09-29T00:25:25 that's the trouble with all these ideas, isn't it? 2010-09-29T00:25:26 that didn't make any sense, we know how to change them, not how to "optimally" update them 2010-09-29T00:25:45 hah http://72.44.46.68/canvas?game_id=214639 2010-09-29T00:25:46 I know what you mean 2010-09-29T00:25:59 not always, still struggling to find ways to do things, let alone refining the heuristics side of it 2010-09-29T00:26:11 well there are a gazillion algorithms out there to search through arbitrary state-spaces 2010-09-29T00:26:45 yeah, but most aren't going to help for this 2010-09-29T00:26:51 you need to reduce the search space a buttload 2010-09-29T00:26:56 never claimed it would :P 2010-09-29T00:27:50 i think i might do some programming this afternoon, now that my test is out of hte way 2010-09-29T00:27:55 i got owned on a linear operators question :( 2010-09-29T00:28:29 *** JCS^ has quit IRC (Quit: \x00) 2010-09-29T00:29:46 meh there are only 4 bots rated higher than mine playing on the tcp server :-\ no wonder my rating is so high 2010-09-29T00:30:14 wtf http://ai-contest.com/visualizer.php?game_id=4939011 :( 2010-09-29T00:30:23 how did i lose that one:-\ 2010-09-29T00:30:53 ah, the timeout of DOOM 2010-09-29T00:31:07 my bot doesn't timeout for sure 2010-09-29T00:31:11 at that level 2010-09-29T00:31:31 nah, server problem, not yours 2010-09-29T00:31:33 there aren't even enough fleets going around 2010-09-29T00:31:38 lots of people having the same problem: http://www.ai-contest.com/forum/viewtopic.php?f=18&t=807 2010-09-29T00:31:58 yet as far as I know timeout is still disqualification in final tournament 2010-09-29T00:32:18 antimatroid: j3camero said it was just a loss 2010-09-29T00:32:20 yesterday 2010-09-29T00:32:33 ah, well i'm glad that has changed 2010-09-29T00:43:50 *** rwa has joined #aichallenge 2010-09-29T00:44:34 how do vectors work in c++? is it a really bad idea to be removing elements from the front of one? 2010-09-29T00:44:43 computation time wise 2010-09-29T00:48:43 antimatroid: I think so, I think you want queue instead 2010-09-29T00:49:04 or some linked list 2010-09-29T00:49:24 "Vector containers are implemented as dynamic arrays; Just as regular arrays, vector containers have their elements stored in contiguous storage locations" 2010-09-29T00:49:29 deque has constant time access to the ends, iirc 2010-09-29T00:49:40 but i could be remembering wrong 2010-09-29T00:49:51 hmmm, a lot of people are just allocating memory for all turns at the start aren't they? 2010-09-29T00:49:59 in haskell i would reach for a finger tree 2010-09-29T00:49:59 double ended queue, yeah 2010-09-29T00:50:03 i would do that, but i'm reluctant to with the number of turns potentially changing 2010-09-29T00:50:29 circular buffers :D 2010-09-29T00:51:07 i would be a big fan of the band of radius for games 2010-09-29T00:51:08 i really hope that we at least get the number of turns specified as part of the game state 2010-09-29T00:51:22 i was hoping we wouldn't but understand why you would want it 2010-09-29T00:51:30 why not? 2010-09-29T00:51:32 a c++ queue uses deque by default, it seems 2010-09-29T00:51:36 I how not too :) 2010-09-29T00:51:47 i don't mean a static number 2010-09-29T00:51:49 cause i don't like players optimising for the end of the game, it should be a battle to the death 2010-09-29T00:51:50 just tell the bots the length 2010-09-29T00:51:56 i mean what you mean, Janzert 2010-09-29T00:52:01 ahh, ok 2010-09-29T00:52:02 antimatroid: but we can't do that 2010-09-29T00:52:14 you could have a band of radius 2010-09-29T00:52:27 antimatroid: without a specified number of turns, what would you say the goal you are optimizing for should be? 2010-09-29T00:52:31 "band of radius"? 2010-09-29T00:52:36 say after t turns, then if distance between growth and ships is greater than x then one player wins 2010-09-29T00:52:46 else? 2010-09-29T00:52:48 otherwise keep playing and reduce the size of x till it gets to like 50 ships or so 2010-09-29T00:52:51 then it's a draw 2010-09-29T00:53:10 that'd be okay maybe, but complicated 2010-09-29T00:53:21 not that complicated 2010-09-29T00:53:35 and leaves the game as one of trying to eliminate the enemy, rather than just beat them on ship coutns 2010-09-29T00:53:45 you'd still end up with the best bots superoptimizing for the endgame condition and it would effectively be the same as having a turn limit, i predict 2010-09-29T00:54:30 *** rwa has quit IRC (Ping timeout: 252 seconds) 2010-09-29T00:54:36 i did like the suggestion on the mailing list that would make games where one player is winning on ship count but the other on growth makes the game a draw 2010-09-29T00:54:36 frankly, i care less about which way these decisions go than actually getting a definitive answer 2010-09-29T00:54:47 *** Oz_ has joined #aichallenge 2010-09-29T00:55:00 hi 2010-09-29T00:55:08 me too 2010-09-29T00:55:17 yeha, i don't think it's fair to call a result otherwise 2010-09-29T00:55:21 *** rwa has joined #aichallenge 2010-09-29T00:55:23 well, except that i want *some* concrete end condition 2010-09-29T00:55:24 *** rwa has quit IRC (Client Quit) 2010-09-29T00:55:50 lol http://72.44.46.68/canvas?game_id=214856 2010-09-29T00:56:31 i mainly just don't want the scenario that j3camero suggested on the mailing list, where we have an entirely unknown endgame condition... 2010-09-29T00:56:38 *almost entirely 2010-09-29T00:56:47 jmcarthur: what was that? 2010-09-29T00:56:52 i am fine with that, provided we know before hand what it is 2010-09-29T00:56:55 intentionally unspecified turn count 2010-09-29T00:57:00 sigh ^^ 2010-09-29T00:57:06 ah 2010-09-29T00:57:16 before hand that that's how it's going to be, that is :P 2010-09-29T00:57:20 antimatroid use a priority Q 2010-09-29T00:57:24 implemented using a min heap 2010-09-29T00:57:25 move 1, DRAW, move 1, DRAW, move 1, DRAW! 2010-09-29T00:57:30 unspecified in the specification but given to the bot in the game is fine with me 2010-09-29T00:57:35 me too 2010-09-29T00:57:40 but that didn't seem to be the proposal 2010-09-29T00:57:41 bhasker: for what? storing future states? 2010-09-29T00:57:43 c++ stl has priority queue 2010-09-29T00:57:54 which is what I thought j3camero meant but maybe not 2010-09-29T00:58:02 well i use it a lot in my bot to prioritize stuff 2010-09-29T00:58:02 yes, i'm using one for potential moves in my naive version 2010-09-29T00:58:03 Top 10 players: dmj111(3756), sequoh(3703), felixcoto(3637), sighbot(3611), davidjliu(3608), Hazard(3579), adumlah(3542), Arekku(3517), _iouri_(3512), bix0r4ever(3512) 2010-09-29T00:58:43 i love being able to just define an operator on a struct and then just throw them into a prique, makes my life so much easier 2010-09-29T00:58:43 ha, I'm catching back up to you sigh 2010-09-29T00:58:45 Janzert: my impression is that a lot of people think that having a turn limit at all is a bad thing and a few people are suggesting that making the turn count an unknown would force people to code as if there is no turn count 2010-09-29T00:58:57 Janzert: but i disagree with that idea 2010-09-29T00:59:09 davidjliu: 3 elo points! 2010-09-29T00:59:16 you can't not have a turn limit, as some games will go infinitely 2010-09-29T00:59:21 not even trivially 2010-09-29T00:59:36 agreed, which is why people are suggesting an unknown turn limit rather than no turn limit 2010-09-29T00:59:45 yep, that's what I want 2010-09-29T00:59:54 and i tihnk the band of radius with that would be best 2010-09-29T01:00:02 cause you could allow longer games when necessary 2010-09-29T01:00:18 we could actually settle this pretty easily by declaring any game which ends without complete dominance to be a draw 2010-09-29T01:00:23 hmm, yeah. There has to be a length limit somehow and I think it is certainly best for the bot to be able to tell what it is 2010-09-29T01:00:27 even if it's 10000 vs 1 on the last turn 2010-09-29T01:00:40 jmcarthur: that's crappy idea 2010-09-29T01:00:48 antimatroid: at least it's not arbitrary 2010-09-29T01:00:50 even then I'd want a bot to be able to tell what the limit is 2010-09-29T01:00:55 band of radius!!! 2010-09-29T01:01:02 Janzert: yeah 2010-09-29T01:01:35 is there any plan to pass that info into bots? 2010-09-29T01:01:36 *** Guest77645 has quit IRC (Ping timeout: 245 seconds) 2010-09-29T01:01:36 and if the length of the games could be significantly longer I think making them all draws would be fine too 2010-09-29T01:01:48 you could even have another inner band saying if you're within this, it's a draw 2010-09-29T01:01:50 meh whatever. i strongly disagree with an unknown turn limit, and only mildly disagree with most other suggestions 2010-09-29T01:02:08 sigh: possibly issue 74 is the most likely candidate at this point 2010-09-29T01:02:26 @url issues 2010-09-29T01:02:26 sigh: issues = http://code.google.com/p/ai-contest/issues/list 2010-09-29T01:02:34 *** xd has joined #aichallenge 2010-09-29T01:02:45 Janzert: thanks 2010-09-29T01:03:00 *** xd is now known as Guest90163 2010-09-29T01:03:11 oh, and i strongly agree with telling the bots about any turn counts 2010-09-29T01:03:35 ahahaha, contestbot sounds like he's tired of posting that link 2010-09-29T01:03:51 hmmmm.. what about if we had a longer turn limit with no bots knowing it and if you make it there without a definitive answer then it's a draw? that way your optimal strategy is to win no matter what time it is, and it's still reasonable to make "long term" moves if that's what would eventually win, as presumably that also keeps you alive? 2010-09-29T01:04:18 but adds a large emphasis on finishing your games fast 2010-09-29T01:04:26 I'm in favor of command line arguments 2010-09-29T01:04:41 I should really respond to that comment 2010-09-29T01:04:42 that will have the least impact on existing bots 2010-09-29T01:05:03 the problem is there are bots known to use command line arguments so they can change there behaviour when run locally 2010-09-29T01:05:06 hmmm depends, mine is set up assuming we wouldn't know such information 2010-09-29T01:05:10 such as creating a log file and such 2010-09-29T01:05:40 all the starter kits that I've seen seem to handle comments correctly 2010-09-29T01:05:42 hmm... yeah, it's a problem if they throw an error because of unknown option 2010-09-29T01:05:45 from the last competition i took that it was best not to assume anything not explicitly stated for these comps 2010-09-29T01:05:55 actually, mine will do that right now 2010-09-29T01:05:59 ignore me, I'm stupid 2010-09-29T01:06:09 and comments have been in the spec from the beginning 2010-09-29T01:06:14 and i really dislike being unsure about game specifics, it's alright for rules to be as general as possible, provided they are definitive from the start 2010-09-29T01:06:29 but now i'm repeating myself, so i'll leave this now :) 2010-09-29T01:06:40 *** McLeopold has joined #aichallenge 2010-09-29T01:09:02 antimatroid: "that way your optimal strategy is to win no matter what time it is" <-- define what you are actually optimizing for there 2010-09-29T01:09:37 the optimal strategy is to be awesome 2010-09-29T01:09:43 optimise for awesomness 2010-09-29T01:10:06 antimatroid: when is it a good idea to sacrifice ships for a higher growth rate if i don't know when the game will end? 2010-09-29T01:10:42 if the long run payoff of claiming that growth is positive 2010-09-29T01:10:45 *** Oz_ has quit IRC (Quit: Page closed) 2010-09-29T01:10:54 antimatroid: but that's not optimal if the turn limit is approaching 2010-09-29T01:11:14 it is irrelevant if that's expected to keep you alive and reaching the end condition is a draw 2010-09-29T01:11:25 and you wouldn't make the move ever if it wasn't expected to keep you alive 2010-09-29T01:11:28 oh, i missed the draw part 2010-09-29T01:11:39 that would be assuming a much longer draw length 2010-09-29T01:11:43 but presumably it could still mean the difference between a win/lose and a draw 2010-09-29T01:11:52 which is significant 2010-09-29T01:11:55 how? 2010-09-29T01:12:14 if making a move means i win whereas not making it means i draw 2010-09-29T01:12:22 the win is still more valuable 2010-09-29T01:12:37 but how do i know that it's actually beneficial? 2010-09-29T01:12:43 yes, but you don't know when the turn limit is to say it will prevent you from winning, same as the other bot 2010-09-29T01:12:55 and it's not going to cause you to lose 2010-09-29T01:13:02 when will i *ever* know to make such a move? 2010-09-29T01:13:07 it adds more draws, but removes unecessary losses 2010-09-29T01:13:17 a draw still counts negatively against a top ranking bot 2010-09-29T01:13:35 scenario: bot is losing, two options: 1) high risk, high reward strategy where fail is quick loss, 2) long drawn out game but sure draw 2010-09-29T01:13:36 not if all top bots would have drawn in that situation 2010-09-29T01:13:55 i don't mean that particular situation is the problem 2010-09-29T01:14:03 the problem is not *knowing* whether you are in that situation 2010-09-29T01:14:09 it screws up your entire strategy 2010-09-29T01:14:09 yeah, okay, it's rather retarded :) 2010-09-29T01:14:10 fair enough 2010-09-29T01:14:47 my number 1 vote would be with band of radius then, otherwise pass the number of turns in as a parameter 2010-09-29T01:15:00 but an upper limit on the number of turns would be desirable 2010-09-29T01:15:07 if that's going to be the case 2010-09-29T01:15:15 you mean a static number? 2010-09-29T01:15:23 like the number of turns is guaranteed to be less than x? 2010-09-29T01:15:24 no, just a maximum that it can be 2010-09-29T01:15:28 yeah 2010-09-29T01:15:43 i don't object to it, but i wonder about why? 2010-09-29T01:15:57 I still like x turns past last neutral takeover. Kinda like 50 moves after last capture or pawn advance in chess. 2010-09-29T01:16:05 well atm i can probably change to allocating room for all future states now 2010-09-29T01:16:12 whereas if that got too large, you couldn't 2010-09-29T01:16:28 if it wasn't specified and you assumed you could do that, your bot might fall over in the finals, which is undesirable for those reasons 2010-09-29T01:16:29 what is too large? 2010-09-29T01:16:30 McLeopold: heh, i would have to redesign my bot from scratch for that one :P 2010-09-29T01:16:48 eh, i guess there would still be an upper bound 2010-09-29T01:16:50 so nevermind 2010-09-29T01:17:04 jmcarthur: you would just have to track the trigger that added more moves. 2010-09-29T01:17:07 sigh: not sure, but there's a limit :P 2010-09-29T01:17:18 *** Snowplan has joined #aichallenge 2010-09-29T01:17:18 uh oh, jmcarthur has got something better than random heuristics 2010-09-29T01:17:20 antimatroid: if we when to such efforts we would also want to specify the amount of ram you have available, etc. 2010-09-29T01:17:36 *** Arthur_ has quit IRC (Ping timeout: 240 seconds) 2010-09-29T01:17:42 *** davidjliu has quit IRC (Quit: Page closed) 2010-09-29T01:17:42 McLeopold: no that wouldn't work for me 2010-09-29T01:17:52 McLeopold: but that there is an upper bound would be fine, regardless 2010-09-29T01:18:16 jmcarthur: what could you be doing that would prevent that? 2010-09-29T01:18:22 McLeopold: i could just assume that the max game length is something like neutral_planets*50 2010-09-29T01:18:28 McLeopold: :) 2010-09-29T01:18:33 not saying 2010-09-29T01:18:47 i want more planets in maps, do you think we'd be able to get that? 2010-09-29T01:18:52 would you guys hate that? 2010-09-29T01:19:03 I would need more CPU time. 2010-09-29T01:19:15 i would need more memory 2010-09-29T01:19:16 guess that is counterproductive to what we've just been talking about aha 2010-09-29T01:19:49 *** Arthur has joined #aichallenge 2010-09-29T01:19:57 jmcarthur: so you have arrays of 200 in your code? 2010-09-29T01:20:10 seriously, how are you using up memory already? 2010-09-29T01:20:14 i'm using like 1.2mb 2010-09-29T01:20:43 is anyone exhausting time limit yet? 2010-09-29T01:20:44 antimatroid: i'm not, and probably won't be using much. i just meant that memory usage will increase more quickly than cpu usage for me 2010-09-29T01:20:56 one could raise turn limit much more with like 0.5 second moves? 2010-09-29T01:21:03 although now that i think about it, they will probably increase at about the same rate 2010-09-29T01:21:20 I've got a bunch of crashes on the new cloud servers. Maybe I'm at my time limit? 2010-09-29T01:21:23 McLeopold: yeah, and the arrays are critical to my idea 2010-09-29T01:21:30 *** iFire has quit IRC (Remote host closed the connection) 2010-09-29T01:21:32 i can't just dynamically resize them or something 2010-09-29T01:21:33 although, bots aren't taking 1 second now and turn limit is made based of time bots are taking, so i guess that's irrelevant 2010-09-29T01:21:57 *** iFire has joined #aichallenge 2010-09-29T01:22:08 jmcarthur: use python? 2010-09-29T01:22:09 i suppose i could extend it for dynamically limited games, but it would be a bit more work for sure 2010-09-29T01:22:15 haskell 2010-09-29T01:22:47 hmmm, i might tell my bot to act retardedly on the main server before the comp if it ever exceeds 0.9 seconds on a move 2010-09-29T01:23:05 jmcarthur: I read the chapter in "Beautiful Code" about haskell. It looks weird. 2010-09-29T01:23:09 lol 2010-09-29T01:23:19 i am not familiar with that writeup 2010-09-29T01:23:29 jmcarthur: I couldn't get my brain around it other than it's suppose to help you write thread safe code. 2010-09-29T01:23:40 oh was it about STM or something? 2010-09-29T01:23:56 I'm looking it up... 2010-09-29T01:23:57 * Janzert wishes the contest was using a dvcs, any dvcs even. 2010-09-29T01:24:13 Janzert: it's on github now, but i'm thinking maybe that was unofficial 2010-09-29T01:24:44 yeah, as far as I know the official development is still supposed to be through svn 2010-09-29T01:24:47 Beautiful Concurrency. He solves the Santa Claus problem. 2010-09-29T01:24:58 McLeopold: ah, i bet it was STM indeed 2010-09-29T01:25:07 that seems familiar now 2010-09-29T01:25:10 i may have read it 2010-09-29T01:25:34 Yep, STM. Gates and Groups. 2010-09-29T01:26:33 * sigh googles "Santa Claus problem" 2010-09-29T01:26:44 i like stm, but i've never really used it in any real programs. i think stm solved the problem it set out to solve, but haskell's parallelism abstractions made the need for explicit thread management a bit obsolete for many uses anyway 2010-09-29T01:26:51 sigh: don't do it. It's a trap! 2010-09-29T01:27:50 *** keynes has quit IRC (Remote host closed the connection) 2010-09-29T01:27:56 It seemed more useful for big enterprise systems where you don't have one programmer. 2010-09-29T01:27:59 so while stm is a very good alternative to most locking strategies in multithreaded programs, it's not such a great alternative when the problem is for parallelism 2010-09-29T01:28:19 hey, this problem sounds familiar :P 2010-09-29T01:28:26 i tend to just use MVars in multithreaded programs 2010-09-29T01:28:38 Hmm... I should rewrite my bot in the map/reduce style... 2010-09-29T01:29:22 my bot is using a library that provides shape polymorphic, regular, multi-dimensional arrays, and if i wanted i could flip a switch and make it all run in glorious parallel 2010-09-29T01:29:30 I may try haskell out some day. Alot of euler fans use it for problems. 2010-09-29T01:29:37 but alas the rules say i may only use one OS thread 2010-09-29T01:29:45 ugh, euler is the worst, IMO 2010-09-29T01:29:54 fun little exercises, but horrible for demoing or learning a language 2010-09-29T01:29:55 Really? why? 2010-09-29T01:30:21 i like euler for the thinking problems, but not for the code or coding 2010-09-29T01:30:41 and of course people tend to use it for the latter more than the former 2010-09-29T01:31:06 i used it to learn maths 2010-09-29T01:31:09 How far have you gotten? 2010-09-29T01:31:10 people learning haskell by solving euler problems usually end up learning the list monad really well and hardly anything else 2010-09-29T01:31:14 me? 2010-09-29T01:31:20 sorry to be a noob but I'm trying to log my bot better. I'm using Java instead of c and c++, just to learn java better. Any good articles you guys can throw me (beside random google search)?? 2010-09-29T01:31:22 anyone... 2010-09-29T01:31:28 and that's basically just doing prolog programming in haskell :P 2010-09-29T01:31:39 I'm at 257 2010-09-29T01:31:42 i forget how far i've gotten. i got bored with it after a while 2010-09-29T01:31:50 I was at 100% for a while, then I just stopped 2010-09-29T01:32:01 about a year ago 2010-09-29T01:32:07 http://logging.apache.org/log4j/1.2/manual.html 2010-09-29T01:32:14 it was eating up my time 2010-09-29T01:32:17 Snowplan: that was for you 2010-09-29T01:33:04 cool thanks... always been worried about ip stealing off chats 2010-09-29T01:33:37 Snowplan: I've usually log a bunch of stuff for each round. Probably half my code is logging right now. 2010-09-29T01:35:23 Snowplan: For the levels, log errors and errors. If your out of sync with the server, warnings. Issue orders as info. Everything else as debug. 2010-09-29T01:35:34 I write rough code and it has done great things. I need to log cause its a dog eat dog world 2010-09-29T01:35:53 cool man thanks 2010-09-29T01:36:09 Snowplan: you lost me with that last sentence... 2010-09-29T01:36:38 i don't normally log unless i'm actively tracking down a bug, then once i've found it i remove all logging again 2010-09-29T01:36:50 well, i'll log errors i guess 2010-09-29T01:36:53 but not much else 2010-09-29T01:37:10 *** Janzert1 has joined #aichallenge 2010-09-29T01:37:12 Well, when I'm on tcp, just errors. All I do is debug now. 2010-09-29T01:37:17 I log most major decisons my bot makes 2010-09-29T01:37:46 *** yasith_ has joined #aichallenge 2010-09-29T01:37:57 *** yasith has quit IRC (Disconnected by services) 2010-09-29T01:38:00 i might log quite a bit in this contest just due to the nature of my algorithm and my desire to observe whether its behavior is desirable, but i normally wouldn't 2010-09-29T01:38:07 *** yasith_ has quit IRC (Client Quit) 2010-09-29T01:38:34 *** yasith has joined #aichallenge 2010-09-29T01:38:44 *** yasith_ has joined #aichallenge 2010-09-29T01:38:45 Yea, I can't really run in debug mode with breakpoints. Logging has saved me lots of time. 2010-09-29T01:39:04 (time I spend on irc...) 2010-09-29T01:39:12 :) 2010-09-29T01:39:19 *** Janzert has quit IRC (Ping timeout: 264 seconds) 2010-09-29T01:39:36 I want to make sure I know exactly what is going on 2010-09-29T01:39:51 I'm also addicted to watching my bot's log as it plays on tcp 2010-09-29T01:40:13 Snowplan: yeah, definitely very useful 2010-09-29T01:41:06 * antimatroid is shocking at debugging efficiently 2010-09-29T01:41:35 antimatroid: you declared something as an int again :P 2010-09-29T01:41:42 :P 2010-09-29T01:42:00 i am in the middle of large changes, the bot probably wouldn't even run atm 2010-09-29T01:42:08 it's certianly not making any moves 2010-09-29T01:45:38 what besides haskell are you all programing in 2010-09-29T01:46:13 c++ 2010-09-29T01:57:52 *** Palmik has joined #aichallenge 2010-09-29T01:58:03 Top 10 players: dmj111(3753), sequoh(3714), sighbot(3635), felixcoto(3622), davidjliu(3612), Hazard(3586), Arekku(3540), adumlah(3527), Rexxar(3496), _iouri_(3493) 2010-09-29T01:58:58 *** keynes has joined #aichallenge 2010-09-29T02:02:54 anyone participating in the 2v2 2010-09-29T02:07:46 *** keynes has quit IRC (Remote host closed the connection) 2010-09-29T02:07:52 2v2? 2010-09-29T02:11:14 *** Olathe has quit IRC (Ping timeout: 265 seconds) 2010-09-29T02:11:23 woot http://72.44.46.68/canvas?game_id=215897 2010-09-29T02:12:31 Cyndre: well done! 2010-09-29T02:12:37 *** Blkt has joined #aichallenge 2010-09-29T02:13:21 http://72.44.46.68/getplayer?player=Cyndre 2010-09-29T02:14:10 *** anoek has quit IRC (Ping timeout: 255 seconds) 2010-09-29T02:15:40 ty 2010-09-29T02:17:33 *** mega1 has joined #aichallenge 2010-09-29T02:30:22 *** troutwine has quit IRC (Ping timeout: 255 seconds) 2010-09-29T02:38:15 *** Naith has joined #aichallenge 2010-09-29T02:39:36 *** yasith_ has quit IRC (Ping timeout: 240 seconds) 2010-09-29T02:40:07 *** yasith_ has joined #aichallenge 2010-09-29T02:40:16 *** yasith has quit IRC (Ping timeout: 276 seconds) 2010-09-29T02:40:34 *** hellman has joined #aichallenge 2010-09-29T02:45:05 *** niknik1200 has joined #aichallenge 2010-09-29T02:46:11 Hi all! I have an question about rules of challenge - can I upload new version of my bot when old is already play? 2010-09-29T02:46:20 yep 2010-09-29T02:46:55 you can keep uploading as many new versions as you want until the deadline 2010-09-29T02:47:07 Thank you, Janzert1 2010-09-29T02:47:16 your elo score resets each time and the old one stops running, but what Janzer1 said 2010-09-29T02:47:22 *** Janzert1 is now known as Janzert 2010-09-29T02:47:49 Thanks. 2010-09-29T02:47:52 is your username referring to nik nik from that kids tv show? 2010-09-29T02:48:01 *** niknik1200 has quit IRC (Client Quit) 2010-09-29T02:48:12 :( 2010-09-29T02:48:52 *** keynes has joined #aichallenge 2010-09-29T02:50:13 *** kjeldahl has joined #aichallenge 2010-09-29T02:52:39 *** McLeopold has quit IRC (Quit: Leaving) 2010-09-29T02:57:17 *** Arthur has quit IRC (Remote host closed the connection) 2010-09-29T02:58:03 Top 10 players: dmj111(3774), sequoh(3738), davidjliu(3628), sighbot(3623), felixcoto(3612), Hazard(3582), _iouri_(3563), adumlah(3538), Arekku(3535), bix0r4ever(3510) 2010-09-29T03:09:28 *** delt0r has joined #aichallenge 2010-09-29T03:14:49 scoreme 2010-09-29T03:14:49 Cyndre, your current score is 46. 2010-09-29T03:18:07 http://ai-contest.com/visualizer.php?game_id=4948561 that was a good battle 2010-09-29T03:25:17 hmm... red had a very strong start, but lacked in defence which you took advantage of well 2010-09-29T03:25:48 scoreme 2010-09-29T03:25:48 sigh, your current score is 92. 2010-09-29T03:26:08 oh god, i hate irc bots 2010-09-29T03:26:21 except contestbot 2010-09-29T03:26:25 !rainbow awesome! 2010-09-29T03:26:25 antimatroid: awesome! 2010-09-29T03:26:33 nice thievery, cyndre 2010-09-29T03:26:37 thief :p 2010-09-29T03:27:02 I've got a score of 92, though! 2010-09-29T03:27:05 noticed you waiting patiently at first and stealing the first planet that red attacked :p 2010-09-29T03:28:05 *** yasith_ has quit IRC (Ping timeout: 265 seconds) 2010-09-29T03:29:34 *** aerique has joined #aichallenge 2010-09-29T03:29:44 ty 2010-09-29T03:29:56 *** Utkarsh has quit IRC (Ping timeout: 240 seconds) 2010-09-29T03:30:02 blue should not have been able to steal that first planet, red could have defended 2010-09-29T03:30:02 antimatroid - cyndrebot only speaks when spoken to. but silently judges you 2010-09-29T03:30:13 yea that was wierd 2010-09-29T03:30:50 yeah sigh 2010-09-29T03:31:12 not sure what's up with them sending all those units to enemy home planet 2010-09-29T03:31:20 not even close distance 2010-09-29T03:31:24 did good defending :) 2010-09-29T03:31:26 it's easier to defend than to attack at that 2010-09-29T03:32:07 don't underestimate rage tactics! 2010-09-29T03:32:18 lol 2010-09-29T03:32:22 yeah, but this one.... 2010-09-29T03:32:45 he actualy does a good job 2010-09-29T03:32:49 you get like 100 spare ships compared to red because red wasted so many units attacking your far away base from the begining. Your main base can just spawn units while taking 2-3 planets while waiting :/ 2010-09-29T03:33:08 close by that would be better 2010-09-29T03:33:17 yeah 2010-09-29T03:33:30 I need more games 2010-09-29T03:33:31 lol 2010-09-29T03:33:32 still, the 9-*Bot series executes it much better, got me a couple of times 2010-09-29T03:33:38 * Sylph nods 2010-09-29T03:33:42 early attack on home 2010-09-29T03:33:48 9-*Bot? 2010-09-29T03:33:59 on tcp 2010-09-29T03:34:06 * being wildcard :P 2010-09-29T03:35:29 *** Utkarsh has joined #aichallenge 2010-09-29T03:35:30 can't find any good examples,, oh well 2010-09-29T03:35:31 *** otherAntimatroid has joined #aichallenge 2010-09-29T03:35:43 http://72.44.46.68/canvas?game_id=216869 2010-09-29T03:36:24 ah 2010-09-29T03:39:11 *** sigh has quit IRC (Remote host closed the connection) 2010-09-29T03:40:09 taking that 76 sized planet really hurt blue 2010-09-29T03:40:23 probably the turning point 2010-09-29T03:40:42 well, also that they only attack one planet at a time :/ 2010-09-29T03:41:04 with more units than necessary, even 2010-09-29T03:41:22 are the game servers 64 bit machines? 2010-09-29T03:44:37 *** yasith_ has joined #aichallenge 2010-09-29T03:45:27 mega1: you awake? 2010-09-29T03:45:35 aerique: yup 2010-09-29T03:45:43 I think the original server was 64bit 2010-09-29T03:45:57 but in this modern distributed world the architectures differ 2010-09-29T03:46:06 not sure exactly how. 2010-09-29T03:46:46 mega1: ok, i've made an sbcl-1.0.42 package for ubuntu hardy but i only have a 32bit machine. would you still like me to file an issue to update sbcl and point them to that package? 2010-09-29T03:48:08 aerique: is it just 'dpkg-buildpackage -rfakeroot' to build it? 2010-09-29T03:49:26 mega1: i did "dpkg -b dir package-name" but it probablt comes down to the same 2010-09-29T03:50:21 I have access to x86-64 so I can build one if you supply me with the ingredients. 2010-09-29T03:51:37 *** jaspervdj has joined #aichallenge 2010-09-29T03:53:18 The latest AMD64 binary is 1.0.37: http://www.sbcl.org/platform-table.html 2010-09-29T03:53:31 sorry, should have been a private message 2010-09-29T03:58:04 Top 10 players: dmj111(3780), sequoh(3726), davidjliu(3639), felixcoto(3619), Hazard(3614), sighbot(3601), _iouri_(3577), adumlah(3544), Arekku(3532), bix0r4ever(3506) 2010-09-29T04:02:20 *** wvdschel has joined #aichallenge 2010-09-29T04:10:32 *** jimmerton has quit IRC (Quit: jimmerton) 2010-09-29T04:18:58 *** Florian95 has joined #aichallenge 2010-09-29T04:23:19 *** yasith has joined #aichallenge 2010-09-29T04:23:41 hey, anyone ? 2010-09-29T04:24:09 My bot crashes after calling pw.FinishTurn() but before starting the next turn 2010-09-29T04:24:10 *** iFire` has joined #aichallenge 2010-09-29T04:24:25 any idea why this happens ? 2010-09-29T04:26:46 *** Naktibalda has joined #aichallenge 2010-09-29T04:27:17 some logs ? 2010-09-29T04:27:36 *** iFire has quit IRC (Ping timeout: 240 seconds) 2010-09-29T04:27:37 I logged 2010-09-29T04:27:45 using fout 2010-09-29T04:27:58 wait, I'll paste them 2010-09-29T04:28:52 Florian95: or do you want the official log ? 2010-09-29T04:29:56 my log: http://tuxv.pastebin.com/itht6LGX 2010-09-29T04:30:23 the official log just says, initializing 2010-09-29T04:30:38 and yes, it crashes on the 1st turn 2010-09-29T04:34:07 Language ? 2010-09-29T04:34:25 C++ 2010-09-29T04:34:32 I cant help :) 2010-09-29T04:34:43 ah, okay :) but thanks anyway 2010-09-29T04:38:45 yasith: your log is not very verbose 2010-09-29T04:39:02 you should describe orders and log time... 2010-09-29T04:39:13 *** delt0r has quit IRC (Quit: Leaving) 2010-09-29T04:39:19 *** delt0r has joined #aichallenge 2010-09-29T04:39:21 edcba: yes, I think i should. I'll add them to the logs now. 2010-09-29T04:39:29 edcba: thanks for the tip btw :) 2010-09-29T04:40:58 *** sigh has joined #aichallenge 2010-09-29T04:42:22 wow, i just found some of the very bad bugs in the bot i put up days ago, i'm surprised it wasn't much worse 2010-09-29T04:45:56 *** keynes has quit IRC (Ping timeout: 264 seconds) 2010-09-29T04:56:24 *** perror has joined #aichallenge 2010-09-29T04:58:04 Top 10 players: dmj111(3777), sequoh(3739), davidjliu(3645), Hazard(3640), felixcoto(3629), sighbot(3594), _iouri_(3578), cfaftw(3578), kcm1700(3541), adumlah(3539) 2010-09-29T05:09:01 *** delt0r has quit IRC (Read error: Operation timed out) 2010-09-29T05:10:27 *** delt0r has joined #aichallenge 2010-09-29T05:22:37 *** Arthur_ has joined #aichallenge 2010-09-29T05:29:38 *** keynes has joined #aichallenge 2010-09-29T05:38:06 lol antimatroid 2010-09-29T05:38:08 happened to me too 2010-09-29T05:38:22 for some reason when I copy pasted stuff around, PlanetOwner turned into PlanetID 2010-09-29T05:38:23 lol 2010-09-29T05:38:43 but fixed it already back then 2010-09-29T05:46:43 *** Als32 has joined #aichallenge 2010-09-29T05:49:07 *** amstan has quit IRC (Ping timeout: 245 seconds) 2010-09-29T05:54:44 *** RainCT has joined #aichallenge 2010-09-29T05:56:28 you know, one of the thing you can do is like send a lot of size 0 fleet... to try to timeout the enemy. @_@ 2010-09-29T05:56:32 i wouldn't do that, obviously 2010-09-29T05:56:36 but kinda funny thinking about that 2010-09-29T05:56:52 fixed 2010-09-29T05:56:56 ah 2010-09-29T05:57:25 fixed as in ? noticed enemy sending 0 unit fleet sitll 2010-09-29T05:57:35 server ignores size 0 fleets 2010-09-29T05:57:40 * Sylph nods 2010-09-29T05:57:47 weird 2010-09-29T05:58:05 maybe i saw a size 8 fleet and thought it was 0 2010-09-29T05:58:06 Top 10 players: dmj111(3785), sequoh(3755), davidjliu(3682), felixcoto(3654), Hazard(3652), _iouri_(3615), sighbot(3604), cfaftw(3597), Arekku(3555), adumlah(3540) 2010-09-29T05:58:13 Sylph: http://ai-contest.com/forums/viewtopic.php?f=19&t=695 2010-09-29T05:58:41 whoa @_@ 2010-09-29T05:59:30 so people actually did that 2010-09-29T06:00:52 *** Utkarsh has quit IRC (Ping timeout: 252 seconds) 2010-09-29T06:07:30 *** Utkarsh has joined #aichallenge 2010-09-29T06:09:44 *** Falconne has joined #aichallenge 2010-09-29T06:10:36 *** Falconne has quit IRC (Client Quit) 2010-09-29T06:12:02 *** Utkarsh has quit IRC (Ping timeout: 245 seconds) 2010-09-29T06:13:26 *** keynes has quit IRC (Remote host closed the connection) 2010-09-29T06:13:59 *** delt0r has quit IRC (Ping timeout: 276 seconds) 2010-09-29T06:18:36 *** keynes has joined #aichallenge 2010-09-29T06:22:05 *** dmj111 has joined #aichallenge 2010-09-29T06:32:00 was it my imagination, or is there a tar ball of results from the tcp server somewhere? 2010-09-29T06:32:14 i scanned the forum topic, and didn't see it. 2010-09-29T06:32:40 * dmj111 was lazy, and hasn't updated his shell script to save the results... 2010-09-29T06:38:52 *** Itkovian has quit IRC (Quit: Itkovian) 2010-09-29T06:41:13 *** iFire` has quit IRC (Read error: Connection reset by peer) 2010-09-29T06:43:10 *** delt0r has joined #aichallenge 2010-09-29T06:50:48 dmj111: there is not, dhartmei can't find anywhere to host it, it is like ~200MB 2010-09-29T06:55:32 :D 2010-09-29T06:55:35 can finally start coding my bot again 2010-09-29T06:56:05 torrent, sigh? :p 2010-09-29T06:56:12 *** hellman has quit IRC (Ping timeout: 245 seconds) 2010-09-29T06:56:17 nm though 2010-09-29T06:56:22 since torrent isn't convenient i guess 2010-09-29T06:56:39 Sylph: don't ask me, ask dhartmei :P 2010-09-29T06:56:41 sigh: makes sense. 2010-09-29T06:56:48 if it's zipped it'll be much smaller though 2010-09-29T06:56:53 glad he is doing it at all. 2010-09-29T06:56:57 I think that was zipped 2010-09-29T06:57:02 seriously @_@ 2010-09-29T06:57:03 that's crazy 2010-09-29T06:57:30 not sure, that's just the figure he quoted 2010-09-29T06:57:55 I could believe it. Those playback strings are huge, and there are a lot of games going. 2010-09-29T06:58:06 Top 10 players: dmj111(3794), sequoh(3755), davidjliu(3677), felixcoto(3641), Hazard(3636), sighbot(3604), _iouri_(3589), cfaftw(3583), Arekku(3559), kcm1700(3544) 2010-09-29T06:59:17 *** Naith has quit IRC (Quit: Naith) 2010-09-29T07:00:19 dmj just gains and gains 2010-09-29T07:01:23 *** iFire has joined #aichallenge 2010-09-29T07:02:14 *** Naith has joined #aichallenge 2010-09-29T07:02:17 lol 2010-09-29T07:02:19 yeah 2010-09-29T07:02:32 me think dmj hax! :P 2010-09-29T07:02:40 Just kidding 2010-09-29T07:04:12 ahhh, but everyone is closing in. time to get back to work. 2010-09-29T07:09:45 lol 2010-09-29T07:12:39 *** Blkt has quit IRC (Read error: Operation timed out) 2010-09-29T07:13:59 *** delt0r has quit IRC (Read error: Operation timed out) 2010-09-29T07:14:26 15th, closing in on that magic top 10 :) 2010-09-29T07:15:32 nice 2010-09-29T07:15:51 *** u1tr0n has joined #aichallenge 2010-09-29T07:15:57 morning all 2010-09-29T07:17:09 Morning 2010-09-29T07:22:44 *** boegel has joined #aichallenge 2010-09-29T07:23:42 *** Zeiris has quit IRC (Ping timeout: 240 seconds) 2010-09-29T07:33:08 *** delt0r has joined #aichallenge 2010-09-29T07:38:09 *** HakanD has joined #aichallenge 2010-09-29T07:39:05 any updates on the time-out issues the server has been showing? 2010-09-29T07:39:16 has a cause been determined yet? 2010-09-29T07:39:34 my guess it that it's due to shared resources (caches, network, ...) 2010-09-29T07:39:43 did you checked the forum? 2010-09-29T07:39:55 there is a topic about it there 2010-09-29T07:39:59 HakanD: I did, I was wondering if more up to date stuff is to be found here :) 2010-09-29T07:40:11 HakanD: yeah, this one: http://ai-contest.com/forum/viewtopic.php?f=18&t=807 2010-09-29T07:41:55 noone around it seems 2010-09-29T07:45:43 *** Naith has quit IRC (Quit: Naith) 2010-09-29T07:54:36 *** yasith__ has joined #aichallenge 2010-09-29T07:55:36 *** dmj111 has left #aichallenge ("ERC Version 5.3 (IRC client for Emacs)") 2010-09-29T07:56:10 dmj111 is using irc from inside emacs o_O 2010-09-29T07:56:50 *** yasith_ has quit IRC (Ping timeout: 240 seconds) 2010-09-29T07:57:01 Emacs is a pretty good operating system, but it could use a better text editor. 2010-09-29T07:57:05 :) 2010-09-29T07:58:07 Top 10 players: dmj111(3753), sequoh(3719), davidjliu(3640), felixcoto(3628), Hazard(3596), _iouri_(3590), sighbot(3576), cfaftw(3575), Arekku(3545), kcm1700(3533) 2010-09-29T07:58:08 *** yasith has quit IRC (Ping timeout: 276 seconds) 2010-09-29T08:04:17 *** delt0r has quit IRC (Read error: Connection reset by peer) 2010-09-29T08:04:29 *** delt0r has joined #aichallenge 2010-09-29T08:04:57 actually Emacs is incredibly powerful-- 2010-09-29T08:05:02 it has a "vi mode" that you can use 2010-09-29T08:05:22 *** Appleman1234 has quit IRC (Ping timeout: 245 seconds) 2010-09-29T08:05:53 *** Appleman1234 has joined #aichallenge 2010-09-29T08:07:54 but does it have a planet-wars-mode? 2010-09-29T08:10:05 *** yasith_ has joined #aichallenge 2010-09-29T08:10:15 I'm pretty sure someone is already working on it 2010-09-29T08:10:29 *** mrSpec has joined #aichallenge 2010-09-29T08:10:32 Hello 2010-09-29T08:20:41 I'd like to take part in this Challange with few friends. I'd like to ask if we can do anything to be sure that we will "fight" with each other? Does teams guarantee that? 2010-09-29T08:21:46 *** BtbN has joined #aichallenge 2010-09-29T08:23:17 Well if you only want to compete within your groups it is possible to download the game engine and viewier and simply pit bots against eachother with that 2010-09-29T08:23:38 The contest itself will probably play you against eachother but it will also play you against everyone else 2010-09-29T08:26:59 pgpaskar_: ok, thanks. So we'll play on server, if it didnt work we'll think about download game engine. Thanks! 2010-09-29T08:32:14 *** Olathe has joined #aichallenge 2010-09-29T08:35:40 *** krokokrusa has joined #aichallenge 2010-09-29T08:35:58 *** delt0r_ has joined #aichallenge 2010-09-29T08:36:03 *** delt0r has quit IRC (Read error: Connection reset by peer) 2010-09-29T08:36:51 haha http://72.44.46.68/canvas?game_id=220789 2010-09-29T08:37:28 mrSpec: you get a lot more games on the tcp server 2010-09-29T08:37:34 you are bound to end up playing each other a bit 2010-09-29T08:39:16 Not sure if it useful to anyone else but I made png's of all the maps with the planet and their id's. I still use the default ShowGame.jar so it's useful for me while debugging: http://www.xs4all.nl/~euqirea/DUMP/map1.png || http://www.xs4all.nl/~euqirea/DUMP/map2png.zip 2010-09-29T08:39:18 *** Utkarsh has joined #aichallenge 2010-09-29T08:40:20 antimatroid: yes, but it we'd like to "use" this game for our "AI in games" subject @ university. Our teacher demands to play with eachother. 2010-09-29T08:42:06 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-09-29T08:42:11 just make somekind cup league and play with locally 2010-09-29T08:42:28 everyone sends it's 'executable' 2010-09-29T08:42:40 *** Nuriaion has joined #aichallenge 2010-09-29T08:43:26 *** mightybyte has joined #aichallenge 2010-09-29T08:45:00 ok :) BTW. as game engine you think about this starter pack, or some other package? 2010-09-29T08:46:19 server's code 2010-09-29T08:46:40 *** marger has joined #aichallenge 2010-09-29T08:48:05 *** Appleman1234 has quit IRC (Excess Flood) 2010-09-29T08:48:15 Naktibalda: Could you give me link to server's code? I cant find it on ai-contest.com :S 2010-09-29T08:48:35 *** Appleman1234 has joined #aichallenge 2010-09-29T08:52:21 tcp server's code is http://www.benzedrine.cx/planetwars/ 2010-09-29T08:52:28 official server - http://code.google.com/p/ai-contest/source/browse/ 2010-09-29T08:53:01 Naktibalda: thanks! 2010-09-29T08:58:07 Top 10 players: dmj111(3747), sequoh(3709), davidjliu(3633), felixcoto(3627), sighbot(3586), _iouri_(3542), kcm1700(3539), Arekku(3531), adumlah(3512), cfaftw(3502) 2010-09-29T09:00:36 *** anoek has joined #aichallenge 2010-09-29T09:02:34 *** Naith has joined #aichallenge 2010-09-29T09:06:25 *** delt0r_ has quit IRC (Read error: Operation timed out) 2010-09-29T09:06:41 Does the compile-server handle FreePascal-Code? 2010-09-29T09:12:54 *** Nuriaion has quit IRC (Quit: Nuriaion) 2010-09-29T09:14:08 no 2010-09-29T09:14:14 but I have seen a ticket for that 2010-09-29T09:14:21 @where issues 2010-09-29T09:14:21 Naktibalda: issues = http://code.google.com/p/ai-contest/issues/list 2010-09-29T09:18:41 *** otherAntimatroid has left #aichallenge 2010-09-29T09:30:57 *** u1tr0n has quit IRC (Ping timeout: 276 seconds) 2010-09-29T09:31:33 *** Appleman1234 has quit IRC (Ping timeout: 264 seconds) 2010-09-29T09:32:59 *** Appleman1234 has joined #aichallenge 2010-09-29T09:33:17 *** keynes has quit IRC (Ping timeout: 252 seconds) 2010-09-29T09:34:13 *** u1tr0n has joined #aichallenge 2010-09-29T09:35:31 *** Appleman1234 has quit IRC (Excess Flood) 2010-09-29T09:36:05 *** Appleman1234 has joined #aichallenge 2010-09-29T09:46:39 hmm, ELO score of bot #10 was at least 200 lower yesterday 2010-09-29T09:58:07 Top 10 players: dmj111(3762), sequoh(3707), davidjliu(3651), felixcoto(3623), sighbot(3586), _iouri_(3549), kcm1700(3538), Arekku(3533), adumlah(3518), cfaftw(3516) 2010-09-29T09:58:22 *** Itkovian has joined #aichallenge 2010-09-29T10:01:33 Naktibalda: I bet that's the time-out bug talking 2010-09-29T10:01:48 was it fixed? 2010-09-29T10:02:55 Naktibalda: it's in the works afaik 2010-09-29T10:03:03 Naktibalda: see http://ai-contest.com/forum/viewtopic.php?f=18&t=807&start=30 2010-09-29T10:03:45 I haven't lost a single game because of timeout today 2010-09-29T10:04:10 probably because I fixed a few bugs related to invalid orders tonigh 2010-09-29T10:04:10 *** lotek93 has quit IRC (Ping timeout: 252 seconds) 2010-09-29T10:04:17 *** delt0r has joined #aichallenge 2010-09-29T10:04:50 Naktibalda: it's a server issue, several people are seeing it 2010-09-29T10:05:05 Naktibalda: I believe someone mentioned it's the game server crashing at times 2010-09-29T10:05:19 Naktibalda: you should check out all the games you lost 2010-09-29T10:05:35 I checked recent games, all correct 2010-09-29T10:05:41 Naktibalda: and see if any ended prematurely? 2010-09-29T10:05:57 Naktibalda: how many games have you played since your latest submission? 2010-09-29T10:06:02 yesterday it was a fault of my bot 2010-09-29T10:06:16 I'm pretty sure my bot is not to blame 2010-09-29T10:06:25 anyway 2010-09-29T10:06:25 afk 2010-09-29T10:06:28 29 2010-09-29T10:07:53 oh, I lost 1 game prematurely, it was a second game 2010-09-29T10:08:01 *** Itkovian has quit IRC (Quit: Itkovian) 2010-09-29T10:08:23 *** wvdschel has quit IRC (Quit: Leaving) 2010-09-29T10:08:34 I won 1 game I was supposed to lose, this way 2010-09-29T10:09:17 I was ashamed of my bot :) 2010-09-29T10:13:38 *** sigh has quit IRC (Remote host closed the connection) 2010-09-29T10:15:12 Im on my up http://ai-contest.com/profile.php?user_id=3911 2010-09-29T10:15:38 *** hellman has joined #aichallenge 2010-09-29T10:15:50 i'm not too far off having a new bot to test out on tcp 2010-09-29T10:15:57 i expect it to be reasonably good 2010-09-29T10:15:58 hmm, it takes a lot of games to climb up to you 2010-09-29T10:16:14 nice 2010-09-29T10:16:19 yea 2010-09-29T10:17:10 oh, you lost against #1317 2010-09-29T10:17:23 it's a leet number 2010-09-29T10:17:46 swiftpants is like 1300 right now 2010-09-29T10:18:12 he didn't submit since 20th 2010-09-29T10:18:51 *** delt0r has quit IRC (Ping timeout: 255 seconds) 2010-09-29T10:20:51 *** Naith has quit IRC (Quit: Naith) 2010-09-29T10:21:05 *** u1tr0n has quit IRC (Read error: Connection reset by peer) 2010-09-29T10:25:09 *** u1tr0n has joined #aichallenge 2010-09-29T10:25:13 *** yasith__ has quit IRC (Ping timeout: 265 seconds) 2010-09-29T10:25:27 *** yasith_ has quit IRC (Ping timeout: 272 seconds) 2010-09-29T10:27:32 *** jesionaj has joined #aichallenge 2010-09-29T10:28:41 *** yasith has joined #aichallenge 2010-09-29T10:36:02 *** HakanD has quit IRC (Ping timeout: 240 seconds) 2010-09-29T10:39:36 *** sugarshark has joined #aichallenge 2010-09-29T10:42:21 *** troutwine has joined #aichallenge 2010-09-29T10:47:09 *** yasith has quit IRC (Ping timeout: 264 seconds) 2010-09-29T10:53:45 *** RainCT has quit IRC (Ping timeout: 264 seconds) 2010-09-29T10:58:08 Top 10 players: dmj111(3754), sequoh(3709), davidjliu(3646), felixcoto(3619), sighbot(3559), kcm1700(3548), _iouri_(3538), cfaftw(3529), Arekku(3515), Imperfectum(3513) 2010-09-29T11:02:10 *** Palmik has quit IRC (Remote host closed the connection) 2010-09-29T11:02:45 *** Palmik has joined #aichallenge 2010-09-29T11:02:49 *** Steps09 has joined #aichallenge 2010-09-29T11:03:55 *** jesionaj has quit IRC (Read error: Connection reset by peer) 2010-09-29T11:04:04 *** Palmik has quit IRC (Remote host closed the connection) 2010-09-29T11:06:22 *** Palmik has joined #aichallenge 2010-09-29T11:09:21 *** RainCT has joined #aichallenge 2010-09-29T11:09:21 *** RainCT has joined #aichallenge 2010-09-29T11:11:15 *** keynes has joined #aichallenge 2010-09-29T11:12:28 *** aerique has quit IRC (Quit: ...) 2010-09-29T11:25:50 *** marger has quit IRC (Quit: Page closed) 2010-09-29T11:32:24 *** perror has quit IRC (Read error: Connection reset by peer) 2010-09-29T11:32:46 *** perror has joined #aichallenge 2010-09-29T11:34:47 *** jesionaj has joined #aichallenge 2010-09-29T11:36:21 *** boegel has quit IRC (Ping timeout: 264 seconds) 2010-09-29T11:39:43 *** u1tr0n has quit IRC () 2010-09-29T11:41:31 mrSpec Once we've got everythign stabled out 2010-09-29T11:42:09 mrSpec We really want to be able to allow player creation of "leages" that play bots against each other. 2010-09-29T11:42:23 *** dvf_ is now known as DanielVF 2010-09-29T11:42:59 DanielVF: it would be cool :) 2010-09-29T11:44:50 DanielVF: Do you have any estimate on how soon you might be pushing changes on the server back into svn? 2010-09-29T11:45:13 Janzert Doing that today, before I start on error coding 2010-09-29T11:45:29 Janzert I should have done it earlier 2010-09-29T11:45:49 ok, thanks. Just a little worried to work on much without knowing what has changed 2010-09-29T11:46:17 Yeah - sorry about holding you up. :( 2010-09-29T11:46:58 no problem, I can certainly understand how it goes 2010-09-29T11:48:02 I've nailed down which cloud server's we'll be using 2010-09-29T11:48:11 great 2010-09-29T11:48:19 Amazon EC2 high CPU medium's. 2010-09-29T11:48:58 That's #7 here: http://ai-contest.com/game_stats.php 2010-09-29T11:49:16 #8 is rackspace's equivelent price point server 2010-09-29T11:49:17 DanielVF: I have no idea what you mean. 2010-09-29T11:49:26 contestbot I don't care 2010-09-29T11:49:27 DanielVF: I have no idea what you mean. 2010-09-29T11:49:29 :P 2010-09-29T11:49:51 But #8 gives occasional timeouts. 2010-09-29T11:49:55 interesting, I've never done it but I've been tempted to use those for distributed testing of an arimaa playing program 2010-09-29T11:50:17 arimaa = chess like game but much harder for computers 2010-09-29T11:50:20 Ah 2010-09-29T11:50:25 Cool 2010-09-29T11:50:38 I found out how to get amazon servers for approx half price 2010-09-29T11:50:47 They let you bid on them 2010-09-29T11:50:51 sweet, the spot pricing? 2010-09-29T11:50:54 yeah 2010-09-29T11:51:02 "spot .... yeah 2010-09-29T11:51:08 For us, works great 2010-09-29T11:51:24 yeah, it does seem perfect for this situation 2010-09-29T11:52:11 Lets us get a $120/m server for $43/m 2010-09-29T11:53:38 wow, that is great. I hadn't ever figured out the monthly cost difference before 2010-09-29T11:54:02 a little more impressive than cents per hour :) 2010-09-29T11:54:17 Yeah 2010-09-29T11:56:39 too bad reserved instances don't cut the spot price at all 2010-09-29T11:58:08 Top 10 players: dmj111(3739), sequoh(3689), davidjliu(3624), felixcoto(3591), sighbot(3557), _iouri_(3534), kcm1700(3525), Arekku(3498), cfaftw(3493), adumlah(3486) 2010-09-29T11:59:43 *** delt0r has joined #aichallenge 2010-09-29T12:01:03 Heh 2010-09-29T12:01:46 *** DanielVF has quit IRC (Quit: DanielVF) 2010-09-29T12:02:06 *** DanielVF has joined #aichallenge 2010-09-29T12:02:19 *** Appleman1234 has quit IRC (Excess Flood) 2010-09-29T12:03:16 *** Appleman1234 has joined #aichallenge 2010-09-29T12:06:13 *** bhasker has quit IRC (Quit: bhasker) 2010-09-29T12:14:41 even if the map didn't include an enemy, and all you had to find was the orders that produced the most ships at the end of the 200th turn, that alone would be NP complete, would you agree? 2010-09-29T12:17:29 hmm, I'm not sure. But I don't think it would follow that a simple strategy that dominates all other strategies to eliminate the enemy is ruled out by that 2010-09-29T12:19:22 gah, it's so tempting to mess around with parameters when you put together a new bot 2010-09-29T12:19:25 although I'm growing more reassured all the time that such a "super strategy" is going to be simple and easy to find 2010-09-29T12:19:25 then it quickly turns messy 2010-09-29T12:19:33 hehe 2010-09-29T12:20:16 err, ...is *not* going to be... 2010-09-29T12:27:10 dhartmei: How do you figure? (with NP hard) 2010-09-29T12:29:33 well, i'm trying to think of a faster algorithm, but it looks hard 2010-09-29T12:29:38 keep in mind NP hard means you can reduce any NP hard problem to it 2010-09-29T12:31:22 *** yasith has joined #aichallenge 2010-09-29T12:31:57 is there a name for this in graph theory? like, you have to pay for edge weights, and start with some value, and get growth? 2010-09-29T12:32:37 i only found stuff about fan arrays and water pump pressure, i must be searching with the wrong terms :) 2010-09-29T12:38:28 *** bhasker has joined #aichallenge 2010-09-29T12:40:31 *** yasith has quit IRC (Ping timeout: 240 seconds) 2010-09-29T12:44:01 *** keynes has quit IRC (Ping timeout: 276 seconds) 2010-09-29T12:44:38 project selection problem? 2010-09-29T12:47:45 minus equipment being shared, looks like it 2010-09-29T12:48:04 *** HakanD has joined #aichallenge 2010-09-29T12:49:31 *** justin_pdx has joined #aichallenge 2010-09-29T12:57:02 *** u1tr0n has joined #aichallenge 2010-09-29T12:58:10 Top 10 players: dmj111(3747), sequoh(3680), davidjliu(3626), felixcoto(3610), sighbot(3571), _iouri_(3554), kcm1700(3534), cfaftw(3518), Arekku(3512), adumlah(3484) 2010-09-29T12:58:31 *** iouri_ has joined #aichallenge 2010-09-29T13:03:21 *** keynes has joined #aichallenge 2010-09-29T13:04:05 dhartmei: max flow? 2010-09-29T13:05:03 i don't really know if there is anything in graph theory close enough to this 2010-09-29T13:06:50 question 2010-09-29T13:06:58 do your bots always draw against themselves 2010-09-29T13:07:20 logically i would expect them to given mirror matchups 2010-09-29T13:07:28 good question! 2010-09-29T13:07:40 I have no random in the bot, but it wins itself on some maps 2010-09-29T13:07:57 it shouldn't 2010-09-29T13:07:59 the only explanation I have is that some maps are not symmetrical... 2010-09-29T13:08:09 but I'm too lazy to check it 2010-09-29T13:08:12 yea i can't understand it either 2010-09-29T13:08:14 they are symmetrical 2010-09-29T13:08:20 and/or a bug 2010-09-29T13:08:21 I mean since distance is calculated as an integer 2010-09-29T13:08:44 it's possible that for some pair of planets it will be +-1 compared to symmetrical pair 2010-09-29T13:08:55 thats possbile 2010-09-29T13:08:58 maybe some rounding poblems 2010-09-29T13:09:03 there are n^2 pairs, so I figure there is a possibility 2010-09-29T13:09:23 maybe rounding problem 2010-09-29T13:09:30 but it exists in the engine too 2010-09-29T13:09:40 of course 2010-09-29T13:09:55 don't you think it should be fixed if so? :) 2010-09-29T13:10:12 it's bad to have advantage for one side of the map, even such miniscule one 2010-09-29T13:10:22 :) 2010-09-29T13:10:42 i think it's more likely that your bot isn't symmetrical 2010-09-29T13:10:54 after all, the planets aren't numbered the same on each side :) 2010-09-29T13:10:54 how it can be ? 2010-09-29T13:10:59 hmm why would my bot not be symmetrical if the map is? 2010-09-29T13:11:03 I ignore numbers :) 2010-09-29T13:11:08 I don't have random 2010-09-29T13:11:17 do you iterate through the list of planets? 2010-09-29T13:11:25 often they are in a certain order 2010-09-29T13:11:30 always sorted by distance or other metric 2010-09-29T13:11:38 which should be symmetrical in essense 2010-09-29T13:11:44 same here 2010-09-29T13:11:49 some planets are equidistant from other planets 2010-09-29T13:12:08 hm, good point 2010-09-29T13:18:03 *** HakanD has quit IRC (Ping timeout: 240 seconds) 2010-09-29T13:19:28 *** Florian95 has quit IRC (Remote host closed the connection) 2010-09-29T13:23:20 *** Naktibalda_ has joined #aichallenge 2010-09-29T13:28:20 *** kjeldahl has quit IRC (Remote host closed the connection) 2010-09-29T13:29:19 *** McLeopold has joined #aichallenge 2010-09-29T13:29:33 *** bhasker_ has joined #aichallenge 2010-09-29T13:30:25 *** Burgundy has joined #aichallenge 2010-09-29T13:32:16 *** bhasker has quit IRC (Ping timeout: 240 seconds) 2010-09-29T13:34:16 *** bhasker has joined #aichallenge 2010-09-29T13:37:44 *** bhasker_ has quit IRC (Ping timeout: 240 seconds) 2010-09-29T13:37:45 there are assymmetries in the maps, but not the way you think. The planet's IDs are what creates a bit of an asymmetry. If two planets (say id=4 and id=5) are equidistant from both players, when sorting planets by distance the planet with id=4 will always appear before planet with id=5 in both players' lists 2010-09-29T13:38:21 but positionally, the first planet might offer advantage to first player while the second planet might offer advantage to the second player 2010-09-29T13:38:56 but because they're equidistant from both players and have the same # of ships and growth rates, both players will go for that planet, and the first player will end up with advantage 2010-09-29T13:39:17 or not 2010-09-29T13:39:32 yep 2010-09-29T13:40:03 first player is in disatvantage, because second captures cheaper 2010-09-29T13:40:06 yeah I missed the point with uncertain ordering in some cases 2010-09-29T13:40:43 Naktibalda_: not when the planets are equidistant 2010-09-29T13:40:55 then they should come at the same turn 2010-09-29T13:40:58 yeah 2010-09-29T13:41:07 no disatvantage for either player 2010-09-29T13:41:08 hm 2010-09-29T13:41:10 true 2010-09-29T13:41:34 ok, that's probably not a good example then 2010-09-29T13:41:44 in ideal draw both players should attack central planet at the same time with same amount of ships and never capture it 2010-09-29T13:41:58 iouri_: I've had bugs due to planet ordering. I've worked through them now. 2010-09-29T13:41:58 ragebot is pretty symmetrical 2010-09-29T13:42:05 and silly to watch 2010-09-29T13:42:07 better example is when both bots have 2 equidistant (or equal by another metric) planets 2010-09-29T13:42:28 1st bot goes for its' 1st planet, 2nd for it's second, creating assymetry 2010-09-29T13:42:35 ahh 2010-09-29T13:42:45 krokokrusa: it's a simetry 2010-09-29T13:42:51 it's mirror symmetry 2010-09-29T13:42:53 hm. I should probably consider that in my bot then 2010-09-29T13:42:59 but maps have central symmetry, not mirror 2010-09-29T13:43:12 hence assymetry ensues 2010-09-29T13:43:26 the problem is that loop (pw.MyPlanets() or pw.Planets() or etc) goes thorigh id's 2010-09-29T13:43:34 yep 2010-09-29T13:44:09 or, emulator can send mirrored coordinates to player 2 :) 2010-09-29T13:44:25 in a way he replaces player id's 2010-09-29T13:44:27 actually, this theory can be tested :) 2010-09-29T13:44:46 you need to re-number all planets for 2nd bot, make decisions, then re-number back when issuing orders :) 2010-09-29T13:45:34 or use more then 1 distance measurement 2010-09-29T13:45:52 ie: first turn check the distance to target planet from your planet and your enemys planet 2010-09-29T13:46:25 *** iouri_ has quit IRC (Quit: Page closed) 2010-09-29T13:46:26 This is what I call symmetrical game http://visualizer.naktibalda.lt/76 2010-09-29T13:47:58 :D 2010-09-29T13:52:54 I have a variable called "enemy_in_range". I typed "enemy_in_rage" by accident. :) 2010-09-29T13:53:18 I came back "True". :) 2010-09-29T13:53:42 both variables look useful 2010-09-29T13:54:34 *** ajhager has quit IRC (Quit: WeeChat 0.3.3) 2010-09-29T13:54:35 *** Bobng has joined #aichallenge 2010-09-29T13:58:10 Top 10 players: dmj111(3763), sequoh(3710), davidjliu(3634), felixcoto(3618), sighbot(3582), _iouri_(3579), kcm1700(3537), Arekku(3537), cfaftw(3534), adumlah(3501) 2010-09-29T13:58:55 I think more useful would be "enemy_out_to_lunch." More amusing, at least. 2010-09-29T14:00:48 *** perror has quit IRC (Quit: Bye all !) 2010-09-29T14:05:17 *** sinan has joined #aichallenge 2010-09-29T14:09:34 *** RainCT has quit IRC (Ping timeout: 245 seconds) 2010-09-29T14:10:24 *** sinan has quit IRC (Ping timeout: 245 seconds) 2010-09-29T14:11:07 *** RainCT has joined #aichallenge 2010-09-29T14:11:09 *** RainCT has joined #aichallenge 2010-09-29T14:14:32 *** bhasker_ has joined #aichallenge 2010-09-29T14:16:44 *** bhasker has quit IRC (Ping timeout: 265 seconds) 2010-09-29T14:19:09 *** hellman has quit IRC (Remote host closed the connection) 2010-09-29T14:22:30 *** bhasker has joined #aichallenge 2010-09-29T14:22:44 http://ai-contest.com/visualizer.php?game_id=4982427 meh 2010-09-29T14:24:23 *** certhia has joined #aichallenge 2010-09-29T14:24:49 *** bhasker_ has quit IRC (Ping timeout: 252 seconds) 2010-09-29T14:24:49 *** certhia has quit IRC (Client Quit) 2010-09-29T14:30:34 *** Palmik has quit IRC (Remote host closed the connection) 2010-09-29T14:38:27 *** krokokrusa has quit IRC (Ping timeout: 252 seconds) 2010-09-29T14:46:33 right, new bots up on main site, time for bed 2010-09-29T14:46:37 only 4 40am :) 2010-09-29T14:50:49 *** boegel has joined #aichallenge 2010-09-29T14:50:56 *** amstan has joined #aichallenge 2010-09-29T14:51:09 *** ChanServ sets mode: +o amstan 2010-09-29T14:52:11 can has win against dmj :) 2010-09-29T14:52:50 Grats! 2010-09-29T14:52:59 Link? 2010-09-29T14:53:05 http://72.44.46.68/canvas?game_id=225825 2010-09-29T14:53:13 we swapped positions aha 2010-09-29T14:53:39 yo y'all 2010-09-29T14:53:48 hello 2010-09-29T14:53:52 any news on the time-outs issues on the main server? any cause found yet? 2010-09-29T14:54:17 no idea sorry 2010-09-29T14:54:17 Yeah I was gonna say, nice flip. You got that extra jump on him that gave you a higher income earlier on. Well done. 2010-09-29T14:57:07 *** iFire has joined #aichallenge 2010-09-29T14:58:11 Top 10 players: dmj111(3770), sequoh(3706), davidjliu(3638), felixcoto(3616), _iouri_(3598), sighbot(3579), Arekku(3547), cfaftw(3542), kcm1700(3542), Mistmanovx(3520) 2010-09-29T14:58:27 *** troutwine has quit IRC (Ping timeout: 240 seconds) 2010-09-29T15:01:02 man i love reducing a function from 10 lines down to just 1 2010-09-29T15:01:14 jmcarthur Yup! 2010-09-29T15:01:48 the way i normally work is i bash out a solution and then clean it up before moving on, so i get to do this a lot 2010-09-29T15:02:03 but it's normally not a 90% reduction 2010-09-29T15:02:34 i go to great lengths to put stuff on one line 2010-09-29T15:02:44 *** area has quit IRC (Ping timeout: 240 seconds) 2010-09-29T15:02:51 to be honest though this is because i removed some extra responsibility from the code and placed it on the client code 2010-09-29T15:02:51 you should see my declarations and instantiations of multidimensional vectors 2010-09-29T15:03:02 how many lines of compiled code was that reduction? it depends on the expressiveness of the language how much goes to cycle saving. also, increased readability is its own benefit 2010-09-29T15:03:09 but i personally prefer giving responsibility to client code anyway 2010-09-29T15:03:41 jmpespxoreax: this was 10 lines of dense haskell reduced to one line of simple haskell 2010-09-29T15:03:46 neat 2010-09-29T15:04:39 cycle saving wasn't even a consideration in this reduction (and rarely is anyway) 2010-09-29T15:05:25 antimatroid: ha, you should see mine :P 2010-09-29T15:06:06 antimatroid: the implementation for this function is one line, but the type signatures is six lines if i remove my type synonyms 2010-09-29T15:06:09 *** troutwine has joined #aichallenge 2010-09-29T15:06:14 *signature 2010-09-29T15:12:06 *** Migi32 has joined #aichallenge 2010-09-29T15:12:18 is the server 32-bit or 64-bit? 2010-09-29T15:12:23 (the main server) 2010-09-29T15:12:29 *** sugarshark has quit IRC (Remote host closed the connection) 2010-09-29T15:12:48 i've heard both, so i have no idea 2010-09-29T15:14:25 ok. Well I guess it's not that important if all maps have less than 32 planets 2010-09-29T15:14:33 (which is true, right?) 2010-09-29T15:15:13 right 2010-09-29T15:15:16 these kinds of things should be in the specification 2010-09-29T15:15:19 do you use 1 bit per planet? 2010-09-29T15:15:44 for some things I'll use bitfields 2010-09-29T15:15:58 or at least, I'll try them out and see how they'll work 2010-09-29T15:16:25 *** als has joined #aichallenge 2010-09-29T15:16:54 also, I have made an awesome compression algorithm that can encode the planet's X,Y position, growth rate, owner and number of ships in one bit 2010-09-29T15:17:01 *** als32 has joined #aichallenge 2010-09-29T15:17:09 ;) 2010-09-29T15:17:25 bit? 2010-09-29T15:17:43 hahaha 2010-09-29T15:18:03 i hope you have a good decompression algorithm, then 2010-09-29T15:18:37 lol 2010-09-29T15:19:05 Zannick, according to my schedule I have to worry about decompression only next week 2010-09-29T15:19:08 first things first 2010-09-29T15:19:13 is there even a memroy limitations 2010-09-29T15:19:47 i mean you'reu nlikely to use much memory anyway 2010-09-29T15:19:54 even if you keep track of everything :| 2010-09-29T15:20:37 *** als has quit IRC (Ping timeout: 252 seconds) 2010-09-29T15:20:41 *** u1tr0n has quit IRC () 2010-09-29T15:22:12 javabots are using >700M :) 2010-09-29T15:22:34 i wouldn't use a 32bit bitfield for planets. i see no reason we can't have more than 32 planets 2010-09-29T15:22:56 and anyway.... what can you actually *do* with a bitfield for planets? .... i guess you can stored whether you own it or not? 2010-09-29T15:23:02 *can store 2010-09-29T15:23:04 http://pastebin.com/E4Zjdcs3 2010-09-29T15:23:11 from the irc logs, stating what CPU the ai-contest server has 2010-09-29T15:23:18 jmcarthur, another reason I use C++ over java is that I can just typedef my bitsets 2010-09-29T15:23:21 for Migi32, who was asking about the bit width 2010-09-29T15:23:40 Migi32: but what's the point? you can only store one bit per planet that way 2010-09-29T15:23:41 javabots are using >700M :) 2010-09-29T15:23:46 yeah i saw that in the forum, pretty ridiculous lol 2010-09-29T15:23:52 jmcarthur: basically any yes or no condition about a planet 2010-09-29T15:24:04 Sylph: its not "real" usage however 2010-09-29T15:24:07 * Sylph nods 2010-09-29T15:24:08 do i own it, do i want it, does the enemy own it, can the enemy take it 2010-09-29T15:24:08 okay 2010-09-29T15:24:20 its addressable vers real vers shared etc 2010-09-29T15:24:22 jmcarthur can't go into that, it's part of the reason I will win this contest ^^ 2010-09-29T15:24:34 lol 2010-09-29T15:24:36 good luck :p 2010-09-29T15:24:41 is it a 5-growth planet, is it a 4-growth planet... 2010-09-29T15:24:45 but to give an answer, it's to store one boolean value per planet 2010-09-29T15:24:47 :) 2010-09-29T15:24:56 Zannick: sure, but as soon as you have 8 of those (not unrealistic) you might as well just store a byte for each planet instead of a word for each property 2010-09-29T15:25:19 sure 2010-09-29T15:25:20 at least my suggestion would scale up with number of planets 2010-09-29T15:25:59 i'm actually only using bitfields for debug conditions, at the moment 2010-09-29T15:26:01 and i really don't think that cache coherency is an issue for these low numbers 2010-09-29T15:26:13 ok so it's 32-bit, making the answer to this question somewhat more important 2010-09-29T15:26:30 when they're doing the finals will they use the same maps? 2010-09-29T15:26:34 Migi32: personally i wouldn't assume the number of planets to be <= 32 2010-09-29T15:26:42 *** sinan has joined #aichallenge 2010-09-29T15:26:55 *** Bobng has quit IRC (Quit: Leaving) 2010-09-29T15:26:59 in fact, i would love to have some larger maps 2010-09-29T15:27:10 i assume the number of planets is <= 64 2010-09-29T15:27:13 even if only for a little fun 2010-09-29T15:27:13 jmcarthur, srsly I need bit-fields to do certain stuff 32 times faster (if not more) than they would with regular bools 2010-09-29T15:27:18 *** Bobng has joined #aichallenge 2010-09-29T15:27:19 just for preallocation 2010-09-29T15:27:23 max 3 planets 2010-09-29T15:27:28 Migi32: i get it, i just don't think it sounds reliable 2010-09-29T15:28:25 is there any "official" of this contest who can give a statement about this (and the 200-move rule, I'd like to know that too) 2010-09-29T15:28:27 Migi32: does stl provide a packed bit vector or anything? 2010-09-29T15:28:37 Migi32: 200 is not specified 2010-09-29T15:28:42 jmcarthur, yeah it's called int 2010-09-29T15:28:45 ;) 2010-09-29T15:28:52 Migi32: (i wrote the specification, and it's still not set in stone) 2010-09-29T15:29:08 Migi32: int has a very small limit :P 2010-09-29T15:29:21 i mean something that can actually store more than 32 or 64 bools 2010-09-29T15:29:40 ah yes, then it's std::bitset or something 2010-09-29T15:29:50 a lot of codes might choke on the competition if you suddenly increase it from 30-40 planets to 200 planets 2010-09-29T15:29:50 lol 2010-09-29T15:29:52 that's what i would use then, personally 2010-09-29T15:29:55 *** Blkt has joined #aichallenge 2010-09-29T15:30:03 i want more planets :( 2010-09-29T15:30:10 or vector but for some reason the people in ##C++ don't like vector 2010-09-29T15:30:12 but agree things needs to be specified from the start 2010-09-29T15:30:17 huh, i wonder why 2010-09-29T15:30:36 vector is specialized for bit packing? nice 2010-09-29T15:30:58 good evening everyone 2010-09-29T15:31:00 interesting 2010-09-29T15:31:06 evening 2010-09-29T15:31:29 Migi32: the turn limit is under active debate, for the record 2010-09-29T15:32:37 if I were running this tournament, I'd just fix it at some value asap, because if you change it late in the tournament it's like we've been training for running a marathon by training sprinting 2010-09-29T15:32:52 Migi32: there is a patch waiting to be merged to include the turn count in the game state reported by the engine as a comment, but i don't think it's been decided whether we will even have a concrete turn limit or something more dynamic anyway... 2010-09-29T15:33:01 Migi32: i agree 2010-09-29T15:33:30 Migi32: i think including it as a comment in the game state is reasonable too though 2010-09-29T15:34:26 well I'm going to assume there are <= 32 planets, if they ever change it I'm out of the tournament 2010-09-29T15:34:26 *** certhia has joined #aichallenge 2010-09-29T15:34:30 :( 2010-09-29T15:34:41 ow Migi32 2010-09-29T15:34:46 so in case any of you still want to win, I'd suggest changing the amount of planets to 33 ;) 2010-09-29T15:34:54 *** Naith has joined #aichallenge 2010-09-29T15:35:00 Migi32: i likewise am assuming that the number of turns will not be determined based on the current state of the game, and my bot will hurt badly if that changes 2010-09-29T15:35:01 bahaha 2010-09-29T15:35:30 doubt that'll change, jmcarthur 2010-09-29T15:35:41 planet size is much more likely to change. 2010-09-29T15:35:42 Sylph: there are a couple wanting that to change, but i doubt it will work either 2010-09-29T15:35:49 * Sylph nods 2010-09-29T15:35:49 no I don't mean it like I'd rage-quit if they'd change it, I'm just going to assume I can pack my booleans in ints 2010-09-29T15:35:49 *** sinan has quit IRC (Ping timeout: 245 seconds) 2010-09-29T15:35:51 planet size? 2010-09-29T15:35:55 oops 2010-09-29T15:35:57 *** certhia has quit IRC (Client Quit) 2010-09-29T15:35:58 i mean planet counts 2010-09-29T15:35:59 sorry 2010-09-29T15:36:02 ah 2010-09-29T15:36:04 got distracted with 2 other things 2010-09-29T15:36:10 *** Accoun has quit IRC () 2010-09-29T15:36:13 still reworking my code foundations :/ 2010-09-29T15:36:18 before my new bot can even move 2010-09-29T15:37:24 * boegel will need to start over, mostly 2010-09-29T15:37:33 yea 2010-09-29T15:37:39 the code i had was largely done like the day i joined 2010-09-29T15:37:44 and finished the day after i joined >.> 2010-09-29T15:38:00 only got to 300 +- 15 2010-09-29T15:38:03 :/ 2010-09-29T15:38:06 me too 2010-09-29T15:38:12 rewriting it now to something proper 2010-09-29T15:38:21 I am rewriting a good framework as well 2010-09-29T15:38:23 ah 2010-09-29T15:38:28 i am always rewriting 2010-09-29T15:38:31 it's the haskell way 2010-09-29T15:38:33 I switched language a few times lol 2010-09-29T15:39:14 I went to java because I wasn't sure how the visualizer I had interpret the bot codes, but now I know so I went back. >.> 2010-09-29T15:39:17 i am writing way too much untested code 2010-09-29T15:39:20 lol 2010-09-29T15:39:21 same 2010-09-29T15:39:57 tempting to log everything 2010-09-29T15:40:22 * jmcarthur exponentiates everything instead 2010-09-29T15:40:27 *** HakanD has joined #aichallenge 2010-09-29T15:40:31 *** Naith has quit IRC (Read error: Connection reset by peer) 2010-09-29T15:40:36 :p 2010-09-29T15:40:45 I also have a problem with testing code 2010-09-29T15:41:19 i'm not used to unit testing :/ 2010-09-29T15:41:22 i would write tests, but it's hard to come up with meaningful properties to test for a game like this 2010-09-29T15:41:29 i'm still going to try i guess 2010-09-29T15:41:31 like tthat'll take my coding time for something this simple that should instantly owrk >.> 2010-09-29T15:41:43 * Sylph nods at jmcarthur 2010-09-29T15:43:12 in theory: "I will write unit tests covering every line of code!" in practice: "Ok done. Let's see how it does. Ok, looks good, next function." 2010-09-29T15:43:17 lol 2010-09-29T15:43:26 yeah 2010-09-29T15:43:54 My problem is that since I don't understand the game well enough, and I'm not sure what I want my bot to do, it's hard to spec out a behavior. 2010-09-29T15:43:54 Migi32: Then there is a bug in the test code... 2010-09-29T15:44:00 i don't care about code coverage in unit tests at all 2010-09-29T15:46:15 any admins in here? 2010-09-29T15:46:20 what I also don't know yet is how to properly debug my code. I can't use gdb or valgrind because they don't work together with the PlayGame.jar 2010-09-29T15:46:30 do you play galcon at all, justin_pdx? 2010-09-29T15:46:40 you should try the flash game if you don't, that'll give you ideas. 2010-09-29T15:46:51 Sylph: I played the flash game on the website a few times. 2010-09-29T15:46:54 ah 2010-09-29T15:46:58 i see 2010-09-29T15:47:14 *** justin_pdx has left #aichallenge 2010-09-29T15:47:15 I wish galcon had symmetrical maps, some maps are just impossible 2010-09-29T15:47:20 *** justin_pdx has joined #aichallenge 2010-09-29T15:47:21 *** justin_pdx has left #aichallenge 2010-09-29T15:47:27 ah yeah 2010-09-29T15:47:36 at least we get symmetrical here, or it'd be messed up 2010-09-29T15:47:43 time to take a walk and clear my head for a bit 2010-09-29T15:47:45 there are like two basic guides now, justin hrm 2010-09-29T15:47:45 *** als32 has quit IRC (Ping timeout: 252 seconds) 2010-09-29T15:47:49 don't remember the links though 2010-09-29T15:47:56 walks are the secret to my success 2010-09-29T15:48:09 and the weather became awesome here over the last couple days, to boot :D 2010-09-29T15:48:19 oh one is at http://mingledcup.dkjones.org/2010/09/google-ai-challenge-strategy-and-tips.html by dmj 2010-09-29T15:48:58 I need to add timing for my logs hrm 2010-09-29T15:49:11 make sure everything runs fast :| 2010-09-29T15:49:41 *** jmcarthur has quit IRC (Read error: Operation timed out) 2010-09-29T15:49:50 *** Accoun has joined #aichallenge 2010-09-29T15:50:27 *** justin_pdx has joined #aichallenge 2010-09-29T15:50:46 wow, none of the top 10 are in this chatroomo? 2010-09-29T15:50:49 chatroom* 2010-09-29T15:50:56 we should team up 2010-09-29T15:50:59 lol 2010-09-29T15:51:08 Their names may be different than their ai challenge handle. 2010-09-29T15:51:17 lol 2010-09-29T15:51:23 tell each other all our genius ideas 2010-09-29T15:51:23 they often are here though 2010-09-29T15:51:24 you first 2010-09-29T15:51:25 some of them have been here from time to time 2010-09-29T15:51:27 like dmj and stuff 2010-09-29T15:51:45 :p Migi32 2010-09-29T15:51:50 It's all about the coding :| 2010-09-29T15:51:52 It's true. If I was in the top ten I'd make sure people knew it by having a matching identifier. 2010-09-29T15:52:09 my bot will be omniscient. 2010-09-29T15:52:17 that's my idea for my 2nd major iteration of bot 2010-09-29T15:52:28 lol, my bot keeps saying that the only winning move is not to play. 2010-09-29T15:52:32 lol 2010-09-29T15:52:40 okay, yours is more omniscient than mine 2010-09-29T15:53:53 don't think i can beat that :D 2010-09-29T15:54:23 lol justin_pdx :D 2010-09-29T15:54:33 how about a game of chess? 2010-09-29T15:54:58 chess is so 80s, Go is the way to go :D 2010-09-29T15:54:59 :p 2010-09-29T15:55:32 good idea though 2010-09-29T15:55:35 *** Itkovian has joined #aichallenge 2010-09-29T15:55:41 like offer game of chess/etc to everyone 2010-09-29T15:55:47 Go is so 3rd century BC 2010-09-29T15:55:48 and let them fight some premade bot 2010-09-29T15:55:57 while you work on your bot 2010-09-29T15:55:58 lol 2010-09-29T15:56:09 :p mightybyte, but its not a short fad. 2010-09-29T15:56:20 it's a lasting classic :D 2010-09-29T15:56:23 oh it would be so awesome if bots could chat and flame each other 2010-09-29T15:56:41 there are probably some rude chatterbot 2010-09-29T15:56:57 though that'll be neat in game 2010-09-29T15:56:59 no I mean if PlanetWars bots could chat while battling 2010-09-29T15:57:01 make it so much more immersive 2010-09-29T15:57:02 yeah 2010-09-29T15:57:36 makes people waste time and cpu cycle being rude there 2010-09-29T15:58:12 Top 10 players: dmj111(3759), sequoh(3703), davidjliu(3643), felixcoto(3610), _iouri_(3603), sighbot(3568), Arekku(3557), kcm1700(3550), Mistmanovx(3515), cfaftw(3510) 2010-09-29T15:59:01 "Your father smells of elderberries" pw.issue_order( 1, 2, 20 ) 2010-09-29T15:59:53 pw.issue_order( castle, knights, cow ) 2010-09-29T16:00:51 "You momma so fat she has a growth_rate of 50" pw.issue_order( 5, 17, 30); 2010-09-29T16:02:25 lol. "Your momma so fat that every fleet I send out winds up in orbit around her house." pw.issue_order( 1, doesn't matter, whatever ) 2010-09-29T16:03:05 *** delt0r_ has joined #aichallenge 2010-09-29T16:05:01 *** delt0r has quit IRC (Ping timeout: 276 seconds) 2010-09-29T16:05:37 *** mightybyte has quit IRC (Quit: leaving) 2010-09-29T16:05:37 "knock knock" 2010-09-29T16:05:41 "who's there?" 2010-09-29T16:05:43 *** Steps09 has quit IRC (Ping timeout: 252 seconds) 2010-09-29T16:06:00 pw.issue_order(3, 11, 2000); 2010-09-29T16:06:16 Hee hee. 2010-09-29T16:06:41 "Why did the chicken cross the road?" 2010-09-29T16:06:48 pw.issue_order( 1, 2, chicken ) 2010-09-29T16:08:44 *** McLeopold has quit IRC (Read error: Connection reset by peer) 2010-09-29T16:10:38 *** aerique has joined #aichallenge 2010-09-29T16:12:56 *** McLeopold has joined #aichallenge 2010-09-29T16:13:47 or just make the player insult on each issue order 2010-09-29T16:14:12 pw.issue_order(1,2,5,rudeComments) 2010-09-29T16:17:12 nah I'd keep the chat to a minimum, like only one message allowed per 3 turns 2010-09-29T16:17:20 otherwise people will abuse it for debugging 2010-09-29T16:18:32 lol true 2010-09-29T16:19:36 *** jaspervdj has quit IRC (Quit: NEVER GONNA GIVE YOU UP NEVER GONNA LET YOU DOWN) 2010-09-29T16:20:11 *** JensTi has joined #aichallenge 2010-09-29T16:22:13 have there ever been draws? 2010-09-29T16:22:26 Migi32: sure 2010-09-29T16:22:29 I mean, at a high level 2010-09-29T16:22:39 Migi32: it included my bot, does that count as high-level? :P 2010-09-29T16:22:39 they're very rare 2010-09-29T16:22:49 ah 2010-09-29T16:22:53 I have a question 2010-09-29T16:23:04 unless you play ragebot against itself or something 2010-09-29T16:23:14 say there's a neutral planet with 4 ships 2010-09-29T16:23:30 and both P1 and P2 send a fleet containing 5 ships, which arrive in the same turn 2010-09-29T16:23:34 who gets the planet? 2010-09-29T16:23:46 neutral 2010-09-29T16:23:49 yeah 2010-09-29T16:24:00 with 0 ships left 2010-09-29T16:24:02 I mean that it remains neutral 2010-09-29T16:24:04 and what happens if the planet was owned by player 1 2010-09-29T16:24:15 P0 can't send fleets 2010-09-29T16:24:15 owner doesn't change 2010-09-29T16:24:19 if ships become 0, control doesn't change 2010-09-29T16:24:29 Naktibalda_: hmmk 2010-09-29T16:24:33 I should fix that then :) 2010-09-29T16:24:46 the number of ships left is the largest force minus the second largest force 2010-09-29T16:24:54 also, if P1 and P2 attack with each 2 ships and P0 has 5 ships, not 1 but 3 ships remain 2010-09-29T16:24:58 Zannick: owner can only change after processing all the fleets that arrive in that particular turn, right? 2010-09-29T16:25:19 Migi32: ah, really? 2010-09-29T16:25:22 like i said, largest force minus second largest force 2010-09-29T16:25:38 where force is the sum of all ships arriving/already there 2010-09-29T16:25:48 s/ships/fleets/ 2010-09-29T16:26:00 Zannick: of one particular player then? 2010-09-29T16:26:10 of one player, yes 2010-09-29T16:26:18 each side loose the same amount of ships 2010-09-29T16:26:52 only if they sent the same number of ships or one of them won 2010-09-29T16:27:14 P0 has 5 ships, P1 sends 3, P2 sends 2. the result is P0 with 2. 2010-09-29T16:27:24 *** RainCT has quit IRC (Read error: Connection reset by peer) 2010-09-29T16:27:36 * boegel had this all wrong then 2010-09-29T16:27:42 what'd be funny is if i submit my new bot and it don't compile coz of mono missing features. -_- 2010-09-29T16:27:48 after wasting so much time on it 2010-09-29T16:27:57 well, actually, I didn't take into account fleets arriving in the same turn correctly 2010-09-29T16:27:58 I just wish they update more readily. :/ 2010-09-29T16:28:51 is Scheme supported currently? 2010-09-29T16:29:09 *** greghaynes has quit IRC (Remote host closed the connection) 2010-09-29T16:29:13 check starter package page 2010-09-29T16:33:22 Naktibalda_: it's not mentioned there, hence my question 2010-09-29T16:33:54 Naktibalda_: there are some forum posts on Scheme though, but it's unclear whether it's supported on the server 2010-09-29T16:34:06 if it's not here, it's not supported 2010-09-29T16:34:21 of course you can try to upload it :) 2010-09-29T16:35:00 yeah, I guess 2010-09-29T16:35:07 it's not for me though, it's for a colleague 2010-09-29T16:35:10 I don't know any Scheme 2010-09-29T16:35:17 hm, looks like yes. 2010-09-29T16:35:18 but I think he'd be great at this contest 2010-09-29T16:35:29 he's a real smart guy, and knows a lot about algorithms and such 2010-09-29T16:35:35 there's an entry for scheme in compile_anything.py 2010-09-29T16:35:38 I hope I can get him to join in :) 2010-09-29T16:35:45 Zannick: ah, great 2010-09-29T16:35:47 boegel: he can rework the haskell starter package easily, I think 2010-09-29T16:35:51 Zannick: where can I find that? 2010-09-29T16:35:52 compile_anything.py is a mess 2010-09-29T16:36:03 i'm upset my patch to clean it up hasn't been accepted :( 2010-09-29T16:36:05 Itkovian: yeah, but the entry step is much bigger then 2010-09-29T16:36:07 http://code.google.com/p/ai-contest/source/browse/trunk/planet_wars/backend/compile_anything.py 2010-09-29T16:36:25 isn't scheme a dialect of List? 2010-09-29T16:36:28 Lisp 2010-09-29T16:37:15 I think it is, yeah 2010-09-29T16:37:25 boegel: starter pack is an easy step compared to a real bot programming 2010-09-29T16:37:27 Bobng: mostly i think people just avoid some of the issues list or are lazy 2010-09-29T16:37:38 contestbot: where sample 2010-09-29T16:37:38 Zannick: I do not know about 'sample', but I do know about these similar topics: 'sampledb' 2010-09-29T16:37:41 contestbot: where sampledb 2010-09-29T16:37:41 Zannick: sampledb = http://csclub.uwaterloo.ca/~amstan/aichallenge/sample.sql 2010-09-29T16:38:09 yeah Zannick, shame :( 2010-09-29T16:38:12 issue for Scheme: http://code.google.com/p/ai-contest/issues/detail?id=45 2010-09-29T16:38:13 okay, the languages table in here says: 2010-09-29T16:38:25 java, C++, C, Python, C#, Haskell 2010-09-29T16:39:27 that's totally wrong, i know people are using javascript and php 2010-09-29T16:39:31 must be out of date 2010-09-29T16:40:22 Bobng: i looked at your patch a while ago and thought it was good. i will see if i can get it checked in later 2010-09-29T16:40:44 its outdated 2010-09-29T16:40:52 *** troutwine has quit IRC (Ping timeout: 240 seconds) 2010-09-29T16:41:07 just check compile_anything.py (: 2010-09-29T16:41:10 Zannick: it might need a little testing though 2010-09-29T16:41:18 not terribly much, and i can probably fix it up 2010-09-29T16:41:25 If its out of date 2010-09-29T16:41:32 then just add entries to the dictionary 2010-09-29T16:41:48 *** jmcarthur has joined #aichallenge 2010-09-29T16:41:52 thats the advantage of it, rather than having lots of if's and the same code 2010-09-29T16:41:56 yeah :) 2010-09-29T16:42:03 it abstracts it out. maybe a little too much abstraction, byt meh 2010-09-29T16:42:12 i like abstraction 2010-09-29T16:42:18 it occurs in a lot of my own python programming 2010-09-29T16:42:28 yeah, thats why I love twisted so much 2010-09-29T16:42:32 boegel: fwiw Common Lisp is supported, but if your colleague is a rabid schemer he won't have anything to do with it :) 2010-09-29T16:42:37 so i'll feel right at home 2010-09-29T16:42:49 but yeah Scheme is part of the Lisp family of languages 2010-09-29T16:42:56 i actually plan on abstracting the game engine sometime later 2010-09-29T16:42:58 Is the test machine running 64 bit or 32 bit? 2010-09-29T16:43:09 *** sinan has joined #aichallenge 2010-09-29T16:43:15 so it doesn't have to be rewritten for every contest 2010-09-29T16:43:19 by that I mean does it have support for psyco or not? 2010-09-29T16:43:22 yeah that would be good 2010-09-29T16:43:25 it's 32 bit 2010-09-29T16:45:02 aerique: I'm not sure how rabid he is, but I'll find out :) 2010-09-29T16:45:45 boegel: he can always use a minor subset of CL and pretend it's Scheme ;-) 2010-09-29T16:46:13 aerique: we'll see :) 2010-09-29T16:47:52 *** BtbN has quit IRC (Quit: Verlassend) 2010-09-29T16:49:30 *** u1tr0n has joined #aichallenge 2010-09-29T16:50:13 what with off server ? all game or win or loss in first round :) 2010-09-29T16:50:24 *** sinan has quit IRC (Ping timeout: 245 seconds) 2010-09-29T16:52:46 *** boegel has quit IRC (Quit: Leaving) 2010-09-29T16:57:15 *** Zeiris has joined #aichallenge 2010-09-29T16:58:09 are you sure that it isn't a fault of your bot? :) 2010-09-29T16:58:13 Top 10 players: dmj111(3776), sequoh(3689), davidjliu(3653), _iouri_(3650), felixcoto(3638), Arekku(3593), sighbot(3582), Mistmanovx(3561), cfaftw(3543), kcm1700(3532) 2010-09-29T16:58:30 I've never lost on my first round... 2010-09-29T16:58:38 whether on java or c# >.> 2010-09-29T16:58:42 hmm, my bot played 4 games in single minute 2010-09-29T16:58:42 dunno on tcp it work fine and time < 1s 2010-09-29T16:59:02 err, in 6 seconds actually 2010-09-29T16:59:08 and i dont belive that my bot can take 315 place 2010-09-29T16:59:09 *** greghaynes has joined #aichallenge 2010-09-29T16:59:34 which bot is yours 2010-09-29T16:59:41 need to make screen ;) 2010-09-29T16:59:58 there are so few C# programmer on the site... 2010-09-29T17:00:05 http://ai-contest.com/profile.php?user_id=9397 2010-09-29T17:00:07 wonder what's up 2010-09-29T17:00:25 now im down to 402 :) 2010-09-29T17:00:25 like everyone use java, C++, or python 2010-09-29T17:00:42 ah 2010-09-29T17:00:45 so c# is the last of 4 main languages 2010-09-29T17:00:53 lol 2010-09-29T17:00:56 hahaha 2010-09-29T17:01:04 i'm tempted to finish my bot really soon and get way up there 2010-09-29T17:01:10 just to add C to that list :P 2010-09-29T17:01:17 lol 2010-09-29T17:01:35 need to add brainf#ck language to list ;) 2010-09-29T17:01:39 lol 2010-09-29T17:01:41 or ook 2010-09-29T17:02:01 malbolge 2010-09-29T17:02:18 i practically guarantee no one will use it, though 2010-09-29T17:02:21 lol 2010-09-29T17:02:50 u1tr0n: Gemini and NickStupich uploaded new versions soon after your game, probably they had fatal bugs 2010-09-29T17:03:29 yea, his games kept getting losses 2010-09-29T17:03:30 all 6 2010-09-29T17:03:33 so it's likely his fault 2010-09-29T17:03:35 nickstupich 2010-09-29T17:04:49 u1tr0n: you lost a few games on second turn :) 2010-09-29T17:05:04 at nignt i made good version of bot but in morning rewrite it with bad one :( this bot didnt win any fight and im back to previos version wich work not so good but work, and now im on 402 place :D 2010-09-29T17:05:14 lol 2010-09-29T17:05:35 yeah i think my fixes a few days ago made my bot about as good as before the rewrite 2010-09-29T17:05:39 same ranks 2010-09-29T17:07:31 athis game is more fun than civ , lost my interest in civ >.> 2010-09-29T17:07:43 so may not even bother playing it, even if its just released not long ago... 2010-09-29T17:07:47 :) 2010-09-29T17:11:38 *** Janzert has quit IRC (Read error: Connection reset by peer) 2010-09-29T17:13:58 *** Janzert has joined #aichallenge 2010-09-29T17:17:18 *** felixcoto has joined #aichallenge 2010-09-29T17:21:35 yea, your code is borked, u1tr0n :p 2010-09-29T17:21:43 but i haven't coded it in php so i don't know 2010-09-29T17:22:06 i know it :( 2010-09-29T17:22:10 was kinda funny to see php starter packages, tbh, but it is a popular language. 2010-09-29T17:22:25 so it makes sense 2010-09-29T17:22:33 what php version do they use anyway ? 2010-09-29T17:22:35 *** mega1 has quit IRC (Ping timeout: 264 seconds) 2010-09-29T17:22:36 and i make bad ideo of bot with bad architecture now i make new one :) 2010-09-29T17:22:39 ah 5.2.4 2010-09-29T17:22:40 not too bad 2010-09-29T17:22:43 5.2.14 2010-09-29T17:22:50 ah 2010-09-29T17:22:58 yeah 2010-09-29T17:22:58 old bot have only 1 war head 2010-09-29T17:23:09 i'm not even done with my framework yet :/ 2010-09-29T17:23:17 still two set of informations to keep track of 2010-09-29T17:23:24 then start work on the actual bot algorithms 2010-09-29T17:23:29 and heuristics xd 2010-09-29T17:24:57 ouh i found whats wrong i trying to save log file 2010-09-29T17:25:03 ah 2010-09-29T17:25:03 lol 2010-09-29T17:25:05 oops 2010-09-29T17:25:06 mb it crash my bot 2010-09-29T17:25:21 yeah i got a debug var that i can just change 2010-09-29T17:25:24 to remove debugging 2010-09-29T17:25:31 which means logging 2010-09-29T17:25:35 *** sigh has joined #aichallenge 2010-09-29T17:25:43 yep i have too 2010-09-29T17:25:50 top 10 in the room, let's be quiet about the strategies. xd 2010-09-29T17:26:07 but forget to change it :) 2010-09-29T17:26:10 ah 2010-09-29T17:26:16 Just kidding 2010-09-29T17:26:17 hi sigh 2010-09-29T17:28:33 whiiiiiiiiii, it work :) 2010-09-29T17:28:43 *** Appleman1234 has quit IRC (Excess Flood) 2010-09-29T17:36:28 great :) 2010-09-29T17:39:29 Sylph: hi 2010-09-29T17:39:35 :) 2010-09-29T17:42:56 http://72.44.46.68/canvas?game_id=228179 :D DRAW vs RageBot :D 2010-09-29T17:45:16 now learn to expand :) 2010-09-29T17:46:00 I like the graph of that one. 2010-09-29T17:46:27 yep; 2010-09-29T17:46:54 im just test new algoritm wich planets attack in first step 2010-09-29T17:47:10 need make die() after it :) 2010-09-29T17:47:20 Yeah weighting potential targets is a tricky process. 2010-09-29T17:47:32 Lots of stuff you can take into account. It's a rabbit hole for sure. 2010-09-29T17:48:48 but you can start simple and improve it gradually 2010-09-29T17:50:07 it's a goal of contest to find the best targets on each turn 2010-09-29T17:50:40 *** Appleman1234 has joined #aichallenge 2010-09-29T17:51:25 *** joeb_ has joined #aichallenge 2010-09-29T17:51:39 agree with you but i try to make first step different from others 2010-09-29T17:52:37 what can you do differently? 2010-09-29T17:52:38 I think that would be a mistake. There is no difference between turn one and turn two. 2010-09-29T17:52:57 In terms of the game state simply being a collection of fleets and planets and factions. 2010-09-29T17:53:00 on turn 1 there aren't any fleets 2010-09-29T17:53:27 I used my new formula to choose targets on the first move 2010-09-29T17:53:39 but I used it for all turns 30 minutes later 2010-09-29T17:53:41 There is a set of fleets it's just empty. It could easily be a state that is found later in the game. 2010-09-29T17:56:51 i said nothing special about the first turn in the spec 2010-09-29T17:57:13 because technically fleets are a part of the map format, and i see no reason why there couldn't be a map with fleets already on the move on the first turn 2010-09-29T17:58:00 a1k0n has mentioned a desire to try a map with a reinforcement fleet arriving later in the game 2010-09-29T17:58:05 as an example 2010-09-29T17:58:12 Top 10 players: dmj111(3769), sequoh(3705), _iouri_(3669), davidjliu(3650), felixcoto(3640), Arekku(3597), sighbot(3577), kcm1700(3545), Mistmanovx(3543), Hazard(3532) 2010-09-29T18:00:24 yep, i don't see anything special on first turn that the bot shouldn't have considered 2010-09-29T18:00:40 myself 2010-09-29T18:01:39 * u1tr0n hate you all and went to delete few millions lines of first step code... 2010-09-29T18:01:59 lol 2010-09-29T18:02:12 Ooh, I'm jealous. My bot is only 50 lines. 2010-09-29T18:02:17 *** sigh has quit IRC (Ping timeout: 240 seconds) 2010-09-29T18:02:18 lol justin 2010-09-29T18:02:26 i am doing special case for first turn because it is when i memoize everything that won't change 2010-09-29T18:02:40 and besides, i know we only have one planet each :) 2010-09-29T18:02:40 my bot got like no lines yet, but I added like 200+ lines already on the other files xd 2010-09-29T18:02:54 but i did take out 200+ overly verbose and cpu wasting lines too 2010-09-29T18:03:01 That's true, I added a bunch of methods to planets. 2010-09-29T18:03:28 that's the only difference really, Zannick 2010-09-29T18:03:39 yeah 2010-09-29T18:03:41 minimizing wasteful counting later down the line. 2010-09-29T18:03:54 oh, and no fleets 2010-09-29T18:03:56 In my model the planets talk with each other first they send out reinforcement requests, then work out what to do with the rest of their ships on their own. 2010-09-29T18:03:56 But you know, if your bot works for first turn, it should work for the rest anyway... 2010-09-29T18:04:04 so it probably isn't a big deal to recount stuff 2010-09-29T18:04:09 that's partially why i'm programming the first turn first 2010-09-29T18:04:13 i may fold that in later 2010-09-29T18:04:15 not good coding, but it don't matter for competition :P 2010-09-29T18:04:22 like this 2010-09-29T18:05:27 Yeah I may not do that, justin. considered that, but meh. 2010-09-29T18:05:43 I'm doing it through simpler method 2010-09-29T18:05:53 like just pulling fleets, rather than requesting them 2010-09-29T18:05:56 It prevents me also from sending too many ships, since the planet knows how many it is capable of sending, that's all encapsulated. 2010-09-29T18:06:14 there are many ways to do that 2010-09-29T18:06:25 *** FlagCapper has quit IRC (Quit: Page closed) 2010-09-29T18:06:54 Hmm, now I wonder if I can have my "bot" simply be a more involved "planet" model. 2010-09-29T18:07:38 probably can 2010-09-29T18:07:46 not probably 2010-09-29T18:07:47 but i mean 2010-09-29T18:08:11 hrm 2010-09-29T18:08:15 Oh well I'm distracting myself from work now. :p I should stop thinking about this. Watching you guys chat it up is entertaining. 2010-09-29T18:08:26 maybe i'l reconsider that again 2010-09-29T18:08:48 nah nm, i prefer an omniscient point of view, from an AI that observe the entire battlefield and the planets 2010-09-29T18:09:12 yeah me too 2010-09-29T18:09:31 my AI observes the entire map over the entire timeline of the game 2010-09-29T18:09:31 not trying to code a game, but win a competition xD 2010-09-29T18:09:36 yeah 2010-09-29T18:09:59 Don't make me extend my future projection time >.> 2010-09-29T18:10:05 :P 2010-09-29T18:10:16 Do you guys think the game is "solveable" 2010-09-29T18:10:37 i suspect it is possible, but i don't think it will be solved for the contest 2010-09-29T18:10:44 probably 2010-09-29T18:11:00 for every configuration of planets? 2010-09-29T18:11:01 i'm actually a little skeptical, to be honest 2010-09-29T18:11:08 the time limit kinda limits that 2010-09-29T18:11:15 makes that unlikely 2010-09-29T18:11:23 if there's a finite number of planet arrangements why not 2010-09-29T18:11:38 not finite anymore once you take into account taking planets 2010-09-29T18:11:40 time limit is the only real limiter 2010-09-29T18:11:40 and retaking them 2010-09-29T18:11:44 not mentioning the number of fleets one can sent 2010-09-29T18:11:54 *** aerique has quit IRC (Quit: ...) 2010-09-29T18:11:58 Sylph: it's finite with memoization over the state of the game 2010-09-29T18:12:04 yeah 2010-09-29T18:12:17 actually, maybe not 2010-09-29T18:12:24 since fleets can grow without bound 2010-09-29T18:12:30 *ship count 2010-09-29T18:12:31 well, bounded by 200 turns 2010-09-29T18:12:35 ah right 2010-09-29T18:12:36 nevermind then 2010-09-29T18:12:44 still going to get crazy though 2010-09-29T18:12:46 (although technically 200 isn't set in stone) 2010-09-29T18:12:54 true 2010-09-29T18:13:04 but just alter the number :D 2010-09-29T18:13:07 if they change it 2010-09-29T18:13:07 lol 2010-09-29T18:13:16 the number of possible orders at each time step is finite 2010-09-29T18:13:22 and if you can project 200 turns into the future, that's a lot still. 2010-09-29T18:13:29 I think it is solvable for example in the simplest case, two starting planets and no neutral planets, you can solve that one pretty quick. 2010-09-29T18:13:35 but then again, so are chess moves 2010-09-29T18:13:41 well, i'm already projecting the entire game, just not every possible game 2010-09-29T18:13:52 justin_pdx: actually i tried that. couldn't solve it for more than 8 moves out, maybe 12 if i let my computer run for a long time 2010-09-29T18:13:55 in that case don't send any unit, justin_pdx :/ 2010-09-29T18:14:04 branching factor is way too huge 2010-09-29T18:14:12 yeah a1kon 2010-09-29T18:14:13 Exactly. Now do the experiment with one neutral planet. 2010-09-29T18:14:24 *** dmy has quit IRC (Quit: Page closed) 2010-09-29T18:14:27 er, i did it with one neutral planet. 2010-09-29T18:14:34 two starting planets, well, there's not much of a game there 2010-09-29T18:14:53 but it's still a good test of feasibility of traversing an entire game tree 2010-09-29T18:15:00 if you can't do it for just two planets... 2010-09-29T18:15:18 branching factor increases by growth rates every turn 2010-09-29T18:15:24 i'm sure that a few simplifications can be made rather than testing every single combination of ships though. 2010-09-29T18:15:27 well 2010-09-29T18:15:31 not quite what i just said 2010-09-29T18:15:36 definitely 2010-09-29T18:15:44 linear programming is still deceptively tempting 2010-09-29T18:15:49 I think the ideal strategy for the game can be determined by analyzing a system of two starting planets and one neutral planet. All rules and strategies should be applicable in that system. 2010-09-29T18:15:50 but i wasn't 100% certain, so i tried it 2010-09-29T18:15:53 but i don't think that would be possible 2010-09-29T18:16:04 I suspect every symmetrical map is a draw, with perfect play. 2010-09-29T18:16:17 it turns out to be very... game-theoretic. predicting your opponent's move is probably key more than anything. 2010-09-29T18:16:20 McLeopold: I think so too. 2010-09-29T18:16:21 McLeopold: maybe a draw over an infinite number of games 2010-09-29T18:16:33 i don't think perfect play is deterministic though 2010-09-29T18:16:40 therefore, it won't be a draw every time 2010-09-29T18:16:45 right, it is not 2010-09-29T18:17:01 If each bot took turns moving, then it would be. 2010-09-29T18:17:12 i've demonstrated to my own satisfaction that it requires making an assumption about your opponent's move before you see it, so there are mutual exclusion events 2010-09-29T18:17:15 McLeopold: right 2010-09-29T18:17:16 What do you mean by "perfect play is deterministic" jmcarthur? 2010-09-29T18:17:20 but yes, if they took turns, then player 1 loses 2010-09-29T18:17:22 it'll be deterministic if the assumptions, etc are the same :/ 2010-09-29T18:17:27 justin_pdx: i think perfect play is a mixed strategy 2010-09-29T18:17:30 so it'll be draw 2010-09-29T18:17:47 since the maps are symmetric xd 2010-09-29T18:17:52 Sylph: nope. perfect play is often going to be a mixed strategy, i suspect 2010-09-29T18:17:52 a1k0n: not so in checkers 2010-09-29T18:18:06 there's no 'perfect play' in that case xd 2010-09-29T18:18:10 perfect in the first place is deterministic 2010-09-29T18:18:16 right, well, i meant in galcon with "simultaneous turns" but with player 1's moves revealed to player 2 2010-09-29T18:18:17 Sylph: not true 2010-09-29T18:18:27 really good play 2010-09-29T18:18:29 Sylph: by that criteria there is no perfect play 2010-09-29T18:18:29 I think the game requires absolutely zero assumptions about your opponents likely moves. 2010-09-29T18:18:35 is not the same as perfect where there's no more improvement to be made 2010-09-29T18:18:39 ever 2010-09-29T18:18:44 "Perfect play" in r/p/s should result in 33% win/loss/draw ratio. 2010-09-29T18:18:50 I think planet wars in the same. 2010-09-29T18:18:50 a1k0n: i think of it in terms of alpha-beta, where you consider what your opponent can do *now* 2010-09-29T18:19:17 yes but you have to marginalize over your opponent's possible moves 2010-09-29T18:19:26 McLeopold: repeated games of rps can have one opponent trying to guess the other opponent's PRNG of course ;) 2010-09-29T18:19:36 Check out my strategy post on close positions. There is some "more perfect play" than others. 2010-09-29T18:19:36 so you end up maximizing your value over that 2010-09-29T18:19:55 well, definitely 2010-09-29T18:20:10 http://ai-contest.com/forum/viewtopic.php?f=17&t=819 2010-09-29T18:20:14 my bot is taking a very game theoretic approach to this, but i don't know how well it will work in practice 2010-09-29T18:21:24 Sylph: if one accepts a mixed strategy as a strategy rather than a single game played according to that strategy, then the mixed strategy can be "perfect" 2010-09-29T18:21:37 as long as you keep the 'quote' in there :D 2010-09-29T18:21:55 Sylph: where "perfect" just means that there is no other strategy you would want to use 2010-09-29T18:22:08 chess games have perfect moves, with every turn calculated to the end. xd 2010-09-29T18:22:18 forgot when they managed that 2010-09-29T18:22:18 Sylph: but that perfection also relies on your opponent being perfect 2010-09-29T18:22:19 *** Blkt has quit IRC (Ping timeout: 252 seconds) 2010-09-29T18:22:24 naturally 2010-09-29T18:22:26 not really actually 2010-09-29T18:22:27 nm 2010-09-29T18:22:34 it doesn't, It pick the 'best' move 2010-09-29T18:22:40 on every single turn based on possible choices to the very end 2010-09-29T18:22:45 Sylph: a different strategy could be better than a "perfect" strategy if you know an imperfection in your opponent 2010-09-29T18:22:46 that's 'perfect' xd 2010-09-29T18:22:54 then its not perfect. :/ 2010-09-29T18:22:58 if it got 'imperfection' 2010-09-29T18:23:01 wth >.< 2010-09-29T18:23:05 you're confusing me 2010-09-29T18:23:23 perfection just has to be defined, is all i'm saying 2010-09-29T18:23:39 Perfection is "not beatable". 2010-09-29T18:23:49 You can draw against it, but that's it. 2010-09-29T18:23:49 justin_pdx: not the case, IMO 2010-09-29T18:23:56 The prisoner's dilemma is an example of a game with "more perfect" strategies with both players making moves at the same time. But winning is more of a meta game. 2010-09-29T18:24:03 justin_pdx: even a perfect bot can be beaten in rock paper scissors 2010-09-29T18:24:14 in a particular instance of the game 2010-09-29T18:24:23 Space Wars is not rock paper scissors. 2010-09-29T18:24:25 You strategy has to beat the majority of other strategies in play. 2010-09-29T18:24:28 we should have a contest where we make bots to play the meta game 2010-09-29T18:24:32 rock paper scissor is all about luck. :| 2010-09-29T18:24:42 how can you do a 'perfect' bot on it >.> 2010-09-29T18:24:43 they go into an irc channel to discuss the bots they're writing for the game 2010-09-29T18:24:58 I understand with perfect move we mean the move that leads to a win when simulated all possible following moves by the player and the oponnent to the end of the game 2010-09-29T18:24:59 Sylph: an imperfect bot would be easily predictable :P 2010-09-29T18:24:59 luck and subjective analysis of opponent moves 2010-09-29T18:25:07 or a bad bot 2010-09-29T18:25:15 don't need to throw out words like 'perfect' in a game of chance. xd 2010-09-29T18:25:22 better vs worse is good enough 2010-09-29T18:25:24 Itkovian: i don't believe that such a move necessarily exist in planet wars 2010-09-29T18:25:28 Space Wars is not a game of chance. 2010-09-29T18:25:28 *exists 2010-09-29T18:25:43 it exist, planet wars isn't about chance 2010-09-29T18:25:50 whether it can be calculated quickly is a different matter 2010-09-29T18:25:54 jmcarthur: that would be my guess too, but having proof would be nice 2010-09-29T18:25:54 justin_pdx: i disagree 2010-09-29T18:26:01 but it's a prisoner's dilemma kind of thing 2010-09-29T18:26:05 For example, tic tac toe is solveable. It is possible to never lose. 2010-09-29T18:26:22 Though I might be wrong 2010-09-29T18:26:23 justin_pdx: it's a deterministic engine, but when there are two equally valuable moves to make that you must choose between, that introduces nondeterminism 2010-09-29T18:26:37 tic tac toe is solvable because the game space is very small 2010-09-29T18:26:43 justin_pdx: true, but planet wars is simultaneous play 2010-09-29T18:26:49 checkers is solved because the game space is relatively small 2010-09-29T18:26:57 Zannick: for a perfect move you;d assume the resources to be unlimited, of course. 2010-09-29T18:27:00 say you have a map with two starting planets and one central neutral planet. you can attempt to take the neutral if your opponent sends his ships to you then you lose. you can send your ships to your opponent but he might send them to you (nash equilibrium) 2010-09-29T18:27:05 thanks to the game being symmetrical, that don't really matter, jmcarthur. 2010-09-29T18:27:12 Sylph: yes it does 2010-09-29T18:27:16 The fact that it exists does not mean you can compute it any time soon 2010-09-29T18:27:18 I think that space wars has equilibrium. 2010-09-29T18:27:27 justin_pdx: over an infinite number of games, i agree 2010-09-29T18:27:28 I suppose the enemy could send to teh same planet or the other planet 2010-09-29T18:27:31 justin_pdx: over a single game, no 2010-09-29T18:28:24 it's easy to construct prisoners' dilemma situations in planet wars, really 2010-09-29T18:28:25 Sylph: consider a map with two home planets and two neutrals, both neutrals being equal distances from both home planets. it's symmetrical, but the game will be necessarily nondeterministic between two perfect players 2010-09-29T18:28:36 a1kon: that 2 planet scenario is solvable. but the solution only garrentees a draw. 2010-09-29T18:28:38 not necessarily hrm 2010-09-29T18:28:43 just turns into a draw 2010-09-29T18:28:45 nope 2010-09-29T18:28:49 whether they take one planet each 2010-09-29T18:28:51 or attack same planet 2010-09-29T18:28:52 McLeopold: right, but there are two equilibria 2010-09-29T18:28:52 it's same thing 2010-09-29T18:28:56 they stay at the same 2010-09-29T18:28:58 depending on the size of the neutral planet anyway 2010-09-29T18:29:02 *** joeb_ has quit IRC (Quit: Page closed) 2010-09-29T18:29:10 'perfect game' don't necessitate winning 2010-09-29T18:29:14 Sylph: consider also the ability to strike a home planet though 2010-09-29T18:29:20 generally any solvable symmetric scenario guarantees a draw :) 2010-09-29T18:29:21 but the moves with least chance to lose 2010-09-29T18:29:21 or defend 2010-09-29T18:29:54 a1kon: as long as the neutral force to takeover is more than the growth rate, it's a draw 2010-09-29T18:29:55 Sylph: i don't believe such single moves (pure strategies) necessarily exist 2010-09-29T18:29:59 * Sylph nods 2010-09-29T18:30:12 Sylph: i don't believe that planet wars will always have a pure, dominant strategy, is what i mean 2010-09-29T18:30:13 it may not, i think it might exist. 2010-09-29T18:30:18 * Sylph nods 2010-09-29T18:30:23 McLeopold: it's a draw even if not, because both players can choose to send all their ships to the neutral on the first turn 2010-09-29T18:30:51 anyway time to make dinner 2010-09-29T18:31:02 a1kon: then if the second player waits 1 turn, they steal the planet. 2010-09-29T18:31:21 yup 2010-09-29T18:31:54 McLeopold: right, but then why wouldn't the first player wait a turn, too? 2010-09-29T18:31:57 hmm, true. yeah i guess the only stable point is attacking directly 2010-09-29T18:32:06 ...or sitting there indefinitely 2010-09-29T18:32:08 Zannick: they would, thus perfect play is a draw 2010-09-29T18:32:33 even attacking directly isn't guaranteed to work, thanks to growth rates and distance 2010-09-29T18:32:39 yeah 2010-09-29T18:32:55 Seriously, look at all my games on map 30. 2010-09-29T18:32:56 well, if you calculate the growth rate and distance 2010-09-29T18:33:04 and the size of the neutral planets 2010-09-29T18:33:13 you can decide whether attacking directly is good or not, at least 2010-09-29T18:33:18 In a one-neutral-planet system the neutral planet is a poison pill, assuming it's inhabited. If it has zero population, then both players should instantly attack it. 2010-09-29T18:33:27 see how many units the enemy will get extra due to your attacking directly xd 2010-09-29T18:33:31 yeah, so there has to be neutrals closer to you than to your opponent for any sort of attack to make sense 2010-09-29T18:33:34 you can still make it neutral :D 2010-09-29T18:33:38 a draw 2010-09-29T18:33:40 mm 2010-09-29T18:33:50 but a perfect game would mean both bots do the same thing :/ 2010-09-29T18:33:58 justin_pdx: yes, neutral force must be greater than growth rate for draw condition 2010-09-29T18:34:14 They wouldn't have to do the same thing, Sylph. 2010-09-29T18:34:39 maybe 2010-09-29T18:34:40 hrm 2010-09-29T18:34:44 For example, a bot in that situation that did nothing but ragebot would still draw. 2010-09-29T18:34:52 * Sylph shouldn't chat so much 2010-09-29T18:34:59 won't get my new bot done if i chat so much -_- 2010-09-29T18:35:11 You can easily create these test maps. 2010-09-29T18:35:24 yeah, but not the 200 turns projection :p 2010-09-29T18:35:27 P 5 5 0 29 4 2010-09-29T18:35:27 P 0 5 1 100 5 2010-09-29T18:35:27 P 10 5 2 100 5 2010-09-29T18:35:30 thee branching get crazy 2010-09-29T18:35:50 that'll probably take so much space and cpu :/ 2010-09-29T18:35:53 beat rage bot on that one... 2010-09-29T18:36:17 lol 2010-09-29T18:36:18 true 2010-09-29T18:36:23 might be fun 2010-09-29T18:36:30 I'll add that map 2010-09-29T18:36:47 I've got 20 or so with simple situations for unit tests. 2010-09-29T18:36:51 custom maps like that are a good way to test edge cases 2010-09-29T18:37:01 yeah 2010-09-29T18:37:07 haha albert's visualizer is hilarious on that map 2010-09-29T18:37:10 What language do you work with, McLeopold 2010-09-29T18:37:16 python 2010-09-29T18:37:43 a1k0n-0.4.1 beats ragebot on that map after 7 planet captures 2010-09-29T18:37:44 my data structures change around so much, it's a good thing it's typeless 2010-09-29T18:38:54 I lose to ragebot 2010-09-29T18:38:56 :| 2010-09-29T18:38:59 lol 2010-09-29T18:39:04 I'm planning to rewrite in javascript. :) 2010-09-29T18:39:04 my old bot anyway 2010-09-29T18:39:22 lol 2010-09-29T18:39:24 crazy 2010-09-29T18:39:30 just kidding 2010-09-29T18:39:36 at least its not ook 2010-09-29T18:40:20 I takes me 31 turns with my current bot. 2010-09-29T18:40:35 I only have the bot I wrote on the day I joined with fixes the day after :/ 2010-09-29T18:40:45 (I might write a bot in whitespace) 2010-09-29T18:40:53 well, working on one from scratch atm 2010-09-29T18:40:56 still working on frameworks :| 2010-09-29T18:40:58 before the bot itself 2010-09-29T18:41:16 Do you have an URL for that visualiser? 2010-09-29T18:41:44 so anyway that map is a perfect example. the equilibrium strategy is a draw, and ragebot uses that strategy, but if you have a predictive model of ragebot's moves you can win 2010-09-29T18:42:00 or rather ragebot doesn't use that strategy 2010-09-29T18:42:06 a1k0n: I don't predict anything. ;) 2010-09-29T18:42:24 i don't either. i guess you just have to be able to exploit a situation. 2010-09-29T18:42:43 ...and have good defence. 2010-09-29T18:43:14 a1k0n: this contest was more fun when you were the top bot. 2010-09-29T18:43:23 so a "perfect" bot would have to wait for its opponent to do something on that map. 2010-09-29T18:43:28 heh. yeah, i agree! 2010-09-29T18:43:53 yea 2010-09-29T18:43:55 Nope: http://ai-contest.com/forum/viewtopic.php?f=17&t=819 2010-09-29T18:44:05 the opponent would waste their units on something and you can steal it 2010-09-29T18:44:13 no, i mean, on the 3-planet map example 2010-09-29T18:44:16 yup 2010-09-29T18:44:16 i know 2010-09-29T18:44:20 still no 2010-09-29T18:44:25 obviously on maps where there are neutrals closer to you than to your enemy then there's a better way to do it 2010-09-29T18:44:32 * Sylph nods 2010-09-29T18:44:42 That's my second map! 2010-09-29T18:44:49 McLeopold: what does your bot do on turn 1 on the 3-planet map? 2010-09-29T18:45:11 P 5 5 0 29 4 2010-09-29T18:45:11 P 0 5 1 100 5 2010-09-29T18:45:11 P 11 5 2 100 5 2010-09-29T18:45:17 many of your opponents are wasteful there, McLeopold 2010-09-29T18:45:32 dumping all their units unnecessarily on their main planet 2010-09-29T18:45:44 much more than they need 2010-09-29T18:45:58 hrm, or attacking you directly >.> 2010-09-29T18:46:05 (i mean the symmetric 3 planet map) 2010-09-29T18:46:10 * Sylph nods 2010-09-29T18:46:38 51 to the center 2010-09-29T18:46:52 *** troutwine has joined #aichallenge 2010-09-29T18:47:25 in that case why can't an opponent send 50+5*diameter+1 to you and diameter/2*5+1 to the center on turn 2? 2010-09-29T18:47:56 then you have to abandon your planet and take theirs 2010-09-29T18:48:06 because that's too many ships? 2010-09-29T18:48:44 depends on how far away they are i guess 2010-09-29T18:48:47 what's diameter? on the small map it is 10? 2010-09-29T18:48:58 on your map it is. i was using a diameter of like 4 2010-09-29T18:49:20 so both planets are 4 from the neutral? 2010-09-29T18:49:28 2 from the neutral. 2010-09-29T18:49:34 hold on... 2010-09-29T18:50:22 20 to the enemy planet, I draw against Ragebot 2010-09-29T18:50:50 given that the neutral is 29 2010-09-29T18:51:00 If it was less than 20, I might do something different 2010-09-29T18:51:14 *** Migi32 has quit IRC (Remote host closed the connection) 2010-09-29T18:51:39 at 19, i win in 13 2010-09-29T18:51:46 how do you draw if you lose 20 ships first? 2010-09-29T18:51:51 oh wait, nvm 2010-09-29T18:52:02 he loses them too 2010-09-29T18:52:04 misread enemy as neutral 2010-09-29T18:52:09 -_- 2010-09-29T18:52:10 Attacking the enemy doesn't lose you ships. Ever. 2010-09-29T18:52:16 oh 2010-09-29T18:52:47 it does if your metric is your ships/total ships and you're behind in ships 2010-09-29T18:52:59 if your metric is yours-his, then no 2010-09-29T18:53:20 a1k0n: oh... I better code that in. 2010-09-29T18:53:24 haha. 2010-09-29T18:53:34 Nice observation, a1k0n. 2010-09-29T18:53:48 haha 2010-09-29T18:53:56 that is why my bot, when ahead, will draw the enemy down with a stream of attacks 2010-09-29T18:54:03 prevents neutral takeovers, etc. 2010-09-29T18:54:19 Mine does that too, but I never thought to not do that when I'm behind. 2010-09-29T18:54:29 This changes everything! 2010-09-29T18:54:33 yw 2010-09-29T18:54:45 I've seen some of the advanced bots have that kind of "I'm ahead, stall 'em" mode. It's interesting to watch them swicth from a growth/defend strategy to an attrition strategy. 2010-09-29T18:54:50 so far each improvement like that has brought me a weaker bot 2010-09-29T18:55:30 maybe you need to fight for a few extra turns before you turn off the hose? 2010-09-29T18:55:30 "improvement" 2010-09-29T18:55:56 let's call it a nonimproving innovation. i have many of those. 2010-09-29T18:56:15 If you watch my official games, it appears my bot does that switch. 2010-09-29T18:56:17 i've basically been cranking up the accuracy with which i forecast a completely fictional future. 2010-09-29T18:56:22 i may have had some of those 2010-09-29T18:56:24 It's actually just gone catatonic. 2010-09-29T18:56:55 It's a bug. 2010-09-29T18:58:13 Top 10 players: dmj111(3771), _iouri_(3698), sequoh(3696), davidjliu(3655), Hazard(3640), felixcoto(3638), Arekku(3608), sighbot(3582), kcm1700(3566), Mistmanovx(3533) 2010-09-29T18:59:04 *** jfrank has joined #aichallenge 2010-09-29T18:59:39 I think a good strategy when ahead is wear down the enemy by attacking each planet with its growth and save the rest to take over neutrals 2010-09-29T18:59:47 I still have to implement it though 2010-09-29T18:59:55 and get ahead to actually test it :-) 2010-09-29T19:00:25 it'sa better strategy to capture enemy planets 2010-09-29T19:00:33 McLeopold: When behind, you need to be aggressive but also save enough ships to take over planets with a single or a few fleets 2010-09-29T19:00:45 *** sigh has joined #aichallenge 2010-09-29T19:00:45 Naktibalda: the win is determined by the number of ships 2010-09-29T19:01:18 well, yes, but if there are neutrals that are easily taken ... 2010-09-29T19:01:24 capturing enemy planet doesn't change difference in number of ships 2010-09-29T19:01:31 neutrals make it worse 2010-09-29T19:01:45 the point is you need to save the extra ships to take a planet, not to steadily wear down the enemy 2010-09-29T19:01:53 Itkovain: I'm implementing that next. 2010-09-29T19:01:54 I think. 2010-09-29T19:02:15 I call it my "anti-dmj" code. 2010-09-29T19:02:29 LOL 2010-09-29T19:02:30 what is anti-dmj? 2010-09-29T19:02:55 *** Olathe has quit IRC (Ping timeout: 272 seconds) 2010-09-29T19:03:08 Try and beat dmj111 2010-09-29T19:03:15 yeah, but how are you doing taht? 2010-09-29T19:03:24 I'm writing code just to deal with his strategy. 2010-09-29T19:03:30 if enemy == dmj then hax() 2010-09-29T19:03:45 sigh: I'm sure you know better than me! 2010-09-29T19:04:04 heh, I'd don't think I've beaten dmj yet 2010-09-29T19:04:09 No, it's more like, if incoming_fleet_size > 300: freak out 2010-09-29T19:04:33 :D 2010-09-29T19:04:47 lol 2010-09-29T19:04:49 hehe 2010-09-29T19:04:50 do you send a small fleets to all directions? 2010-09-29T19:05:01 Never: that's a waste. 2010-09-29T19:05:07 I'll call that the cactus maneuver :d 2010-09-29T19:05:08 Most games are lost that way. 2010-09-29T19:05:09 :D 2010-09-29T19:05:43 that's probably seen much more at rank 1000+, McLeopold 2010-09-29T19:05:54 nm, more like 1500+ 2010-09-29T19:05:56 maybe 2010-09-29T19:07:08 hmm... I expand too fast to counter dmj :( 2010-09-29T19:07:10 http://ai-contest.com/visualizer.php?game_id=4996814 2010-09-29T19:07:19 at turn 58, I waste all my fleets 2010-09-29T19:07:34 *** Olathe has joined #aichallenge 2010-09-29T19:08:32 *** Apophis has quit IRC (Ping timeout: 255 seconds) 2010-09-29T19:09:00 woah, yeah... taking that neutral was stupid 2010-09-29T19:09:09 you don't check for that? 2010-09-29T19:09:26 that bot is 3 weeks old. 2010-09-29T19:09:34 :) 2010-09-29T19:09:41 and still in the top 20 2010-09-29T19:09:46 your next iteration will be scary :) 2010-09-29T19:10:00 I keep thinking I'll submit it... 2010-09-29T19:10:18 But I want it to punch everyone in the face, so I keep waiting. 2010-09-29T19:10:25 haha 2010-09-29T19:10:27 Until it's "perfect" 2010-09-29T19:11:14 "Everybody's got a plan, until they get punched in the face." 2010-09-29T19:13:02 *** troutwine has quit IRC (Ping timeout: 265 seconds) 2010-09-29T19:13:19 *** JensTi has quit IRC (Quit: Leaving.) 2010-09-29T19:13:49 *** Apophis has joined #aichallenge 2010-09-29T19:19:57 *** Bobng has quit IRC (Quit: Leaving) 2010-09-29T19:23:33 *** McLeopold has quit IRC (Quit: Leaving) 2010-09-29T19:23:37 I wish I could tell if my ranking is significant in any way. 2010-09-29T19:23:50 I mean, I wonder what % of the submissions are just the default package. 2010-09-29T19:23:57 *** Zeiris has quit IRC (Ping timeout: 240 seconds) 2010-09-29T19:23:57 it 2010-09-29T19:24:04 if it's not #1, it doesn't matter 2010-09-29T19:24:10 Oh pssh. 2010-09-29T19:24:12 ;) 2010-09-29T19:24:47 Like, am I losing to altered bots or unaltered bots? What % of altered bots do I win against? 2010-09-29T19:25:32 If you're above the midline, I'm sure you're playing against altered bots 2010-09-29T19:25:40 I'm currently 664th, in case that helps you give me feedback. 2010-09-29T19:26:58 it's pretty easy to tell if you are playing an altered bot, just see if they have more than 3 fleets at a time :P 2010-09-29T19:30:50 i am slowly climbing towards the top 100 ... but taking its own sweet time 2010-09-29T19:31:09 http://ai-contest.com/profile.php?user_id=5822 2010-09-29T19:31:50 sigh i have a question, do you implement look ahead and if so how, one thing i have been considering is my heuristic based approach works fine in some, but in some the bot just gets stupid and selects the wrong starting 2010-09-29T19:31:51 planets 2010-09-29T19:31:58 and then its just downhill from there 2010-09-29T19:32:20 so either a) i need to come up with different heuristics for different variations of the maps or implement some sort of lookahead 2010-09-29T19:32:35 well, my bots knows the future states of the planets based on current ships, but that's about it 2010-09-29T19:32:43 yea that my bot does too 2010-09-29T19:32:47 it doesn't do any searching of the game tree 2010-09-29T19:33:00 neither does mine but i am wondering if its good to do it at start 2010-09-29T19:33:04 like in the first move 2010-09-29T19:33:11 identify all available planets 2010-09-29T19:33:24 and look ahead 5-6 turns and see which maximizes your ships 2010-09-29T19:33:52 I just haven't figured out a good way to do it 2010-09-29T19:34:11 hmm 2010-09-29T19:35:43 there's a lot of stuff I keep trying and failing, and a lot of other stuff I want to do, but don't know how yet 2010-09-29T19:36:51 heh 2010-09-29T19:38:06 i think my current heuristic is a bit better than before 2010-09-29T19:38:07 http://72.44.46.68/canvas?game_id=229848 2010-09-29T19:38:18 but it still gets kinda stupid in maps when starting positions are close by 2010-09-29T19:38:53 *** davidjliu has joined #aichallenge 2010-09-29T19:39:12 *** Appleman1234 has quit IRC (Quit: Leaving) 2010-09-29T19:48:25 *** Apophis has quit IRC (Read error: Connection reset by peer) 2010-09-29T19:50:26 *** Apophis has joined #aichallenge 2010-09-29T19:55:53 *** FlagCapper has joined #aichallenge 2010-09-29T19:56:13 Is timing out randomly on the first move a problem with the server or is it my fault? 2010-09-29T19:57:21 I don't think anyone is completely sure at the moment 2010-09-29T19:57:56 Well if it's happening to other people I won't get too worried about it. 2010-09-29T19:58:13 Top 10 players: dmj111(3738), sequoh(3695), Hazard(3692), _iouri_(3639), davidjliu(3630), felixcoto(3627), Arekku(3579), sighbot(3566), kcm1700(3560), adumlah(3534) 2010-09-29T19:59:47 I know some that were timing out found it to be a problem in the bot, but I don't think server problems have been ruled out for sure as well 2010-09-29T20:01:27 there's often a reason for it, like a noncompiled program being unable to run due to incompatibility 2010-09-29T20:04:08 Well my program runs perfectly fine on my computer, and 1/5th of the games on the server 2010-09-29T20:04:23 Sorry, 4/5ths 2010-09-29T20:04:31 It timeouts randomly occasionally 2010-09-29T20:04:40 *** justin_pdx has quit IRC (Quit: justin_pdx) 2010-09-29T20:04:42 *** felixcoto has quit IRC (Quit: Page closed) 2010-09-29T20:05:51 probably a freak error, then. 2010-09-29T20:06:18 On my part or thiers? 2010-09-29T20:06:48 either, i don't know 2010-09-29T20:07:03 Ok, thanks 2010-09-29T20:07:14 when my bot times out, it usually does it two games in a row or something 2010-09-29T20:07:48 The oddest part is I have my bot set to quit at 1/40th of a second 2010-09-29T20:07:54 I can't possibly see how it would timeout 2010-09-29T20:07:55 when we get error indications, we might be able to narrow it down further 2010-09-29T20:08:20 if it's a server issue, then it's either not letting you run, or not reading "go" correctly 2010-09-29T20:10:03 i was actually testing a bug report yesterday and i found that i needed to flush the output stream in order to not timeout from python on the first turn 2010-09-29T20:14:54 I'm using C++, so I don't know if that matters 2010-09-29T20:15:10 you should still flush stdout at the end of your turn 2010-09-29T20:15:39 I thought the starter bot already did if you're building off that but I may be remembering wrong 2010-09-29T20:15:49 It does 2010-09-29T20:16:08 the starterbot flushes after each order 2010-09-29T20:16:16 and the "go" 2010-09-29T20:16:21 doesn't hurt :) 2010-09-29T20:16:58 well technically I suppose it could take a little bit of unnecessary time 2010-09-29T20:18:01 *** Mathnerd314 has joined #aichallenge 2010-09-29T20:19:41 *** Arthur_ has quit IRC (Ping timeout: 276 seconds) 2010-09-29T20:20:46 *** Janzert has left #aichallenge 2010-09-29T20:20:57 *** jfrank has quit IRC (Quit: Page closed) 2010-09-29T20:21:09 *** Arthur has joined #aichallenge 2010-09-29T20:28:16 hmm, I'm getting "Invalid order" rather often in TCP server, but I don't see anything wrong in my orders 2010-09-29T20:34:40 I want to submit new bot to test out if my C# code will compile on the server... but don't want to lose all my previous fights:/ 2010-09-29T20:34:53 wish it don't like delete everything, just the ranking or something 2010-09-29T20:35:04 Sylph: so... keep the version that you have on the server 2010-09-29T20:35:06 upload thins one 2010-09-29T20:35:11 then upload the other one back 2010-09-29T20:35:15 yeah 2010-09-29T20:35:21 but lose all the current fights 2010-09-29T20:35:28 can't watch them anymore coz they get removed from my page xd 2010-09-29T20:36:13 yeah, I find that annoying too... it would be nice to have a compile test 2010-09-29T20:36:32 Sylph: i'm pretty sure the games don't get deleted 2010-09-29T20:36:39 Sylph: save your profile page 2010-09-29T20:36:48 Sylph: the links should still work after that 2010-09-29T20:37:05 they don't no 2010-09-29T20:37:14 i guess I can like 'save the page 2010-09-29T20:37:27 and use the saves pages to review 2010-09-29T20:37:37 saved 2010-09-29T20:53:03 *** Arthur has quit IRC (Disconnected by services) 2010-09-29T20:53:25 *** Arthur_ has joined #aichallenge 2010-09-29T20:58:13 Top 10 players: Hazard(3768), dmj111(3753), sequoh(3707), _iouri_(3649), davidjliu(3644), felixcoto(3633), Arekku(3587), sighbot(3576), kcm1700(3559), adumlah(3537) 2010-09-29T20:58:58 and Hazard has reached first place! 2010-09-29T20:59:45 dmj's rein in over 2010-09-29T21:00:17 ha, he might be back soon... 2010-09-29T21:00:32 *** krokokrusa has joined #aichallenge 2010-09-29T21:08:19 I'm back 2010-09-29T21:08:23 "you should still flush stdout at the end of your turn" 2010-09-29T21:08:34 Yeah, if you don't do this, your bot won't play 2010-09-29T21:09:04 C standard lib, and most standard IO libs buffer output on the bot side 2010-09-29T21:09:16 until you've generated 4K of output 2010-09-29T21:09:42 So flush on go, or die. 2010-09-29T21:12:10 Question: Will you be disqualified for timeouts in the final tournie? 2010-09-29T21:12:40 s/tournie/tourney 2010-09-29T21:12:55 *** AlliedEnvy has quit IRC (Ping timeout: 245 seconds) 2010-09-29T21:16:25 You'll probably lose that game. 2010-09-29T21:16:56 So exactly how it is now? 2010-09-29T21:17:08 Like that 2010-09-29T21:17:14 Because I know in the Tron contest it was a disqualification. 2010-09-29T21:17:15 Ok 2010-09-29T21:18:16 It's possible that later on we could do something like timeout or crash X games in a row, and stop playing that submission 2010-09-29T21:20:14 maybe for first move crash/timeout for really broken bots, but otherwise it should just be a loss I suggest 2010-09-29T21:20:35 sigh Yeah 2010-09-29T21:20:57 Gonna try to make a distintion between first move timeouts and later ones 2010-09-29T21:21:06 cool, thanks :) 2010-09-29T21:21:37 so glad that's no longer a rule 2010-09-29T21:22:54 *** bhasker has quit IRC (Ping timeout: 276 seconds) 2010-09-29T21:25:34 *** AlliedEnvy has joined #aichallenge 2010-09-29T21:26:02 *** blbrown_win3 has joined #aichallenge 2010-09-29T21:26:15 "Note: OctaneLang.java uses unchecked or unsafe operations." Will this cause a compile fail? Interesting 2010-09-29T21:27:00 *** Naktibalda_ has quit IRC (Remote host closed the connection) 2010-09-29T21:28:36 *** krokokrusa has quit IRC (Ping timeout: 272 seconds) 2010-09-29T21:34:16 yep 2010-09-29T21:42:20 *** Zaphus has joined #aichallenge 2010-09-29T21:58:14 Top 10 players: dmj111(3735), sequoh(3719), _iouri_(3673), Hazard(3657), davidjliu(3649), felixcoto(3639), sighbot(3573), Arekku(3566), adumlah(3542), kcm1700(3537) 2010-09-29T22:08:54 dmj is back on top! 2010-09-29T22:09:09 yea biatches, 496 and 4/4 losses 2010-09-29T22:09:33 http://ai-contest.com/profile.php?user_id=6230 2010-09-29T22:09:35 woot 2010-09-29T22:09:59 lies, I only see 3 losses 2010-09-29T22:10:43 *** troutwine has joined #aichallenge 2010-09-29T22:18:52 eventing all 2010-09-29T22:21:10 Hello. 2010-09-29T22:21:47 scoreme 2010-09-29T22:21:47 Cyndre, your current score is 69. 2010-09-29T22:25:13 oh i remember this 2010-09-29T22:25:42 Ok, I've found out it's basically always my first game that I timeout on the first turn on. 2010-09-29T22:26:11 Or any of the really fast games that happen at the beginning 2010-09-29T22:32:32 scoreme 2010-09-29T22:32:32 Sylph, your current score is 55. 2010-09-29T22:32:39 your bot hates me :/ 2010-09-29T22:33:13 only your first game? weird, flag. 2010-09-29T22:34:23 I'm just going to go ahead and blame the server, easiest way out. 2010-09-29T22:34:49 though not the most productive 2010-09-29T22:34:50 :p 2010-09-29T22:35:00 lol, my bot doesnt hate you, it just hates how you use words 2010-09-29T22:35:05 Yup 2010-09-29T22:35:10 scoreme 2010-09-29T22:35:10 FlagCapper, your current score is 66. 2010-09-29T22:35:20 What does it like ? :) 2010-09-29T22:35:24 a lot of smiley faces ? :) 2010-09-29T22:35:40 What does it score exactly? 2010-09-29T22:35:42 not really sure anymore 2010-09-29T22:35:47 lol 2010-09-29T22:35:51 jmcarthur: its alot better then it was 2010-09-29T22:35:52 been adding words? 2010-09-29T22:37:18 FlagCapper: it scores word usage. Break sentence into word pairs, rank each word pair based on users score, user ranked by current word pair ranks, and if new randomly choose a score and mod based on user score 2010-09-29T22:37:56 *** Zeiris has joined #aichallenge 2010-09-29T22:38:09 Sylph: it does it automaticly 2010-09-29T22:38:12 explainword president 2010-09-29T22:38:12 president = verb: 1 noun: 5 adjective: 1 adverb: 1 pronoun: 0 preposition: 0 person: 92 place: 18 thing: 0 idea: 0 2010-09-29T22:38:12 what makes any word pair any good then. >.> 2010-09-29T22:38:18 Sylph usage 2010-09-29T22:38:25 ok 2010-09-29T22:38:40 it counts number of usage, not the rank of the users that use those word pair, I see 2010-09-29T22:38:45 watch what happens when I use words badly 2010-09-29T22:38:47 scoreme 2010-09-29T22:38:47 Cyndre, your current score is 86. 2010-09-29T22:38:52 lol 2010-09-29T22:38:58 scoreme 2010-09-29T22:38:58 Sylph, your current score is 30. 2010-09-29T22:39:01 crap 2010-09-29T22:39:03 scoreme 2010-09-29T22:39:03 Sylph, your current score is 14. 2010-09-29T22:39:06 crap 2010-09-29T22:39:07 crap 2010-09-29T22:39:08 scoreme 2010-09-29T22:39:08 Sylph, your current score is 7. 2010-09-29T22:39:10 :| 2010-09-29T22:39:14 That's pretty bad 2010-09-29T22:39:23 scoreme 2010-09-29T22:39:23 Sylph, your current score is 38. 2010-09-29T22:39:24 I sylph killed crap current your is score 2010-09-29T22:39:27 scoreme 2010-09-29T22:39:27 Cyndre, your current score is 85. 2010-09-29T22:40:08 my bot 2010-09-29T22:40:10 scoreme 2010-09-29T22:40:10 Sylph, your current score is 51. 2010-09-29T22:40:26 if serious a 2010-09-29T22:40:27 scoreme 2010-09-29T22:40:27 Cyndre, your current score is 154. 2010-09-29T22:40:35 lol 2010-09-29T22:40:43 lol 2010-09-29T22:40:44 if serious a 2010-09-29T22:40:46 scoreme 2010-09-29T22:40:46 Sylph, your current score is 84. 2010-09-29T22:41:14 but even though you used that to get ahead it wont work forever - every time the word ranks puts you ahead, the words are ranked down by your score 2010-09-29T22:41:28 Yeah I know 2010-09-29T22:42:05 I'm going a whole different direction on my code and it now confuses me :| 2010-09-29T22:42:07 hrm 2010-09-29T22:42:20 that sucks 2010-09-29T22:42:21 scoreme 2010-09-29T22:42:22 sigh, your current score is 99. 2010-09-29T22:42:32 askbot who is the president of the united states of america 2010-09-29T22:42:33 obama , people , were , has , help , their , over , does , doesn , black , person , also , us , change , because , world , thats , make , american , still , ti 2010-09-29T22:42:42 well, just starting on the bot side of the code, to be honest. 2010-09-29T22:42:48 finished the framework, 2010-09-29T22:42:51 awesome, I'm almost at 100 2010-09-29T22:42:55 potential answers/subjects to create a sentence from :) 2010-09-29T22:42:58 score sigh 2010-09-29T22:42:59 grats sigh 2010-09-29T22:43:10 @score sigh 2010-09-29T22:43:11 Cyndre: You have no gotten any error messages recently, so here's a random one just to let you know that we care. 2010-09-29T22:43:20 @score Cyndre 2010-09-29T22:43:21 Sylph: User error -- Replace user. 2010-09-29T22:43:21 Has anyone really been far even as decided to use even go want to do look more like? 2010-09-29T22:43:23 scoreme 2010-09-29T22:43:23 jmcarthur, your current score is 151. 2010-09-29T22:43:26 :) 2010-09-29T22:43:47 why are you two so high? 2010-09-29T22:43:55 and that is what my bot likes now sylph 2010-09-29T22:44:07 jmcarthur 2010-09-29T22:44:15 sigh: Because I has anyone really been far even as decided to use even go want to do look more like? 2010-09-29T22:44:28 duh 2010-09-29T22:44:29 wow 2010-09-29T22:44:36 that hurt to read 2010-09-29T22:44:38 ah, of course 2010-09-29T22:44:48 Your bot my bot all around the american black bots because that change is still over 2010-09-29T22:44:50 score 2010-09-29T22:44:55 scoreme 2010-09-29T22:44:55 Sylph, your current score is 30. 2010-09-29T22:44:58 ouch 2010-09-29T22:45:05 funny bot 2010-09-29T22:45:13 its neat cause you can teach it things 2010-09-29T22:45:33 askbot zipcodeppp beats up Sylph 2010-09-29T22:45:35 could , other , after , times , sure , same , hi , two , planet , play , when , map , who , m , minus , got , his , dont , than , does , ve , been , 1 , 2010-09-29T22:45:45 zipcodeppp beats up Sylph with big sticks 2010-09-29T22:45:47 askbot zipcodeppp beats up Sylph 2010-09-29T22:45:48 big , sticks , could , other , after , times , sure , same , hi , two , planet , play , when , map , who , m , minus , got , his , dont , than , does , ve 2010-09-29T22:46:05 just so much noise I used a unique word to demonstrate 2010-09-29T22:46:13 ah 2010-09-29T22:46:19 I need to prune the connections in the memeory 2010-09-29T22:46:25 askbot Hask anyone really 2010-09-29T22:46:26 those , don , had , want , does , dont , has , minus , yeah , up , also , use , when , here , good , more , better , m , could , even , because , need , he 2010-09-29T22:46:45 oops 2010-09-29T22:46:47 oh well 2010-09-29T22:46:52 and like a human - you put in garbage and garbage comes out :) 2010-09-29T22:47:01 askbot Has anyone really 2010-09-29T22:47:03 should , when , any , dont , people , been , much , want , who , which , here , than , his , use , them , im , other , lot , had , very , us , go , else , 2010-09-29T22:47:09 noise :( 2010-09-29T22:47:20 askbot has anyone really flown a plane? 2010-09-29T22:47:21 them , had , down , ever , way , cool , someone , off , okay , into , kludge , should , when , any , dont , people , much , who , which , here , than , use 2010-09-29T22:47:56 askbot what are the primary colors 2010-09-29T22:47:57 half , second , any , switch , still , school , into , way , many , got , has , those , who , brush , 1 , image , sense , different , put , white , mode , 2 2010-09-29T22:48:06 primary colors are red blue and yellow 2010-09-29T22:48:08 askbot what are the primary colors 2010-09-29T22:48:08 half , second , any , switch , red , still , school , into , yellow , blue , way , many , got , has , those , who , brush , 1 , image , sense , different , p 2010-09-29T22:48:34 wtf, spam attack! 2010-09-29T22:48:35 and for it to know something you have to teach it, or have it parse a website 2010-09-29T22:48:37 * sigh hides 2010-09-29T22:49:44 askbot 1 + 1 2010-09-29T22:49:46 2 , 0 , equal , minus , 1divided , 3 , divided , x , times , 6 , 5 , up , 4 , h , y , c , sqrt , 000 , say , between , 15 , person , 9 , r , really , 2010-09-29T22:50:01 askbot sqrt(2) 2010-09-29T22:50:01 equal , 1 , plus , x , divided , minus , 4 , 3 , 1divided , 0 , times , bys , m , 5 , r , lim , got , want , say , right , u , e , has , where , v , 2010-09-29T22:50:25 *** JCS^ has joined #aichallenge 2010-09-29T22:50:44 askbot sigh 2010-09-29T22:50:44 distress , real , right , arent , meant , n , could , tackle , big , guard , got , moved , religion , same , religious , time , against , creature , expression , 2010-09-29T22:52:20 askbot spam sigh back to the channel 2010-09-29T22:52:21 up , go , when , make , time , now , d , put , anyone , has , want , few , ask , very , don , divided , good , did , still , against , ll , here , could , 2010-09-29T22:53:25 FlagCapper: no actual thought - just memory lookups 2010-09-29T22:54:35 askbot alphabet 2010-09-29T22:54:36 q , p , r , o , v , u , z , n , x , w , y , m , l , k , j , h , g , f , e , d , c , b , letters , letter , which , into , teach , modern , minu 2010-09-29T22:55:06 missed "a" but its probably not shown do to overusage 2010-09-29T22:55:16 no I either 2010-09-29T22:57:15 wow my entire terminal is filled with lists of words 2010-09-29T22:58:15 Top 10 players: Hazard(3703), dmj111(3697), sequoh(3687), _iouri_(3656), davidjliu(3627), felixcoto(3621), sighbot(3560), Arekku(3553), adumlah(3532), Mistmanovx(3520) 2010-09-29T22:58:34 *** hellman has joined #aichallenge 2010-09-29T22:59:31 problem is teaching it things - I would love to get a k-12 transcript and feed it to the bot so it just knows what the primary colors are 2010-09-29T23:00:27 lol jmcarthur 2010-09-29T23:00:42 askbot the game 2010-09-29T23:00:42 into , only , play , video , make , time , am , should , m , has , trying , want , them , back , right , ll , same , server , use , who , here , logged , w 2010-09-29T23:01:46 askbot planet wars 2010-09-29T23:01:46 really , same , should , always , don , other , any , pages , exist , strategy , mean , believe , has , under , yes , star , people , over , d , make , now , 2010-09-29T23:01:59 askbot server 2010-09-29T23:01:59 why , 1 , client , irc , game , any , since , has , use , count , restarted , help , 0 , free , highest , see , connection , clients , been , per , now , own 2010-09-29T23:02:33 askbot president missile white house 2010-09-29T23:02:33 take , might , more , still , very , around , order , nothing , come , give , want , use , did , make , people , has , man , im , when , had , bush , were , 2010-09-29T23:02:38 aw CyndreBot_ doesn't respond to /msg 2010-09-29T23:02:40 *** Zaphus has quit IRC (Quit: Page closed) 2010-09-29T23:02:50 jmcarthur: I can make him respond in /msg 2010-09-29T23:03:01 make it him her 2010-09-29T23:03:04 meh whatever. i was just going to try avoiding flooding the channel 2010-09-29T23:03:13 i should be productive anyway 2010-09-29T23:03:21 I did add $sex = 'male'; but not sure people are willing to accept that its a him 2010-09-29T23:03:35 I think they are waiting for it to get the surgery 2010-09-29T23:03:36 lol 2010-09-29T23:04:03 askbot planetwars strategy planet wars 2010-09-29T23:04:03 mean , should , sylph , really , always , good , two , don , game , over , m , other , pure , believe , has , dominant , could , matter , make , than , more , 2010-09-29T23:04:53 but just join #cyndrebot 2010-09-29T23:11:52 guys.. wtf? 2010-09-29T23:12:03 Cyndre: can you not spam like that please? 2010-09-29T23:12:26 yea cause it was interfering with the people interacting with it 2010-09-29T23:12:33 leavechannel 2010-09-29T23:12:33 *** CyndreBot_ has left #aichallenge 2010-09-29T23:13:38 and wtf - was that really necessary when I invited people out of the channel already? 2010-09-29T23:13:58 askbot poop 2010-09-29T23:13:59 Cyndre: sorry, lol, just surprised me 2010-09-29T23:14:10 pfft 2010-09-29T23:17:24 *** hellman has quit IRC (Remote host closed the connection) 2010-09-29T23:18:58 *** Burgundy has quit IRC () 2010-09-29T23:21:55 *** u1tr0n has quit IRC () 2010-09-29T23:28:56 *** Accoun has quit IRC (Ping timeout: 240 seconds) 2010-09-29T23:29:10 *** Accoun has joined #aichallenge 2010-09-29T23:30:04 *** hellman has joined #aichallenge 2010-09-29T23:31:48 *** wh1teside has quit IRC (Ping timeout: 255 seconds) 2010-09-29T23:39:17 *** wh1teside has joined #aichallenge 2010-09-29T23:45:20 *** yasith_ has joined #aichallenge 2010-09-29T23:50:37 *** bhasker has joined #aichallenge 2010-09-29T23:58:15 Top 10 players: Hazard(3738), dmj111(3691), sequoh(3683), _iouri_(3658), felixcoto(3628), davidjliu(3623), sighbot(3568), Arekku(3552), adumlah(3542), Mistmanovx(3525)