2010-11-22T00:00:50 meh, people should just have a cup of concrete 2010-11-22T00:00:52 it requires basically no effort to do.. and doesn't hurt anyone who is entering the 'contest proper' 2010-11-22T00:01:33 does it? are there any FPSes that do that? I know SC2 has leagues, and if you're in the top X of the league you can expect to be put into better league.... don't think it's needed in this case :) 2010-11-22T00:03:19 *** Prillicy has quit IRC (Quit: I be bangin') 2010-11-22T00:03:20 virdo.. I can easily see lots of people, taking the default bot.. making a small change.. seeing it in teh bottom 20%, thinking about it for an hour or two, and giving up 2010-11-22T00:03:43 if there was a bit of a stepping stone, as they worked on their bots.. it'd be much easier to enter 2010-11-22T00:06:53 like, I'm sure EVERY bot has "what is the best lanet to attack right now?" as part of it's logic 2010-11-22T00:08:24 I guess, it would really be no different than teh already proposed, allowing custom groups/ladders 2010-11-22T00:08:59 except, to have a couple specific ones, that slightly modify the rules of the game 2010-11-22T00:14:15 i consider taking many planets at each node 2010-11-22T00:14:20 not just "the best one now" 2010-11-22T00:18:23 so you just have a yes/no.. no metric atlal on which one is best? 2010-11-22T00:18:41 Top 10 players: bocsimacko(3359), dmj111(3196), Raschi(3148), asavis(3145), _Astek_(3107), bix0r4ever(3100), deccan(3086), felixcoto(3085), ChrisH(3056), george(3045) 2010-11-22T00:21:37 *** delt0r___ has joined #aichallenge 2010-11-22T00:23:17 *** delt0r has quit IRC (Ping timeout: 255 seconds) 2010-11-22T00:23:42 *** Appleman1234 has quit IRC (Read error: Connection reset by peer) 2010-11-22T00:24:44 *** Appleman1234 has joined #aichallenge 2010-11-22T00:25:23 burny: how about a contest that prevents all programmers from entering, only allowing computer scientists to enter? and also, no heuristics allowed, the standard AI techniques must be used 2010-11-22T00:26:08 (many CSes seem to turn into programmers) 2010-11-22T00:26:12 2010-11-22T00:26:32 lol 2010-11-22T00:30:32 if anything i'd want more leaderboards for harder maps 2010-11-22T00:30:49 or more players per map 2010-11-22T00:30:50 ya, per map board could be cool too 2010-11-22T00:31:14 I'll get a per map ranking on zeroviz later 2010-11-22T00:32:03 McLeopold_: I'm going to get two different revisions of my bot going next week 2010-11-22T00:32:15 once to handle huge distances and plant numbers, another to handle multiple players 2010-11-22T00:32:19 one* 2010-11-22T00:32:31 nice, zeroviz will support you 2010-11-22T00:32:54 far_crazy will have large distance 2010-11-22T00:33:08 hundreds of planets too please :) 2010-11-22T00:33:16 crazy_dense 2010-11-22T00:34:05 you've played crazy_fast maps, right? 2010-11-22T00:34:05 i'll try to do a minimax revision for small n-player maps 2010-11-22T00:34:13 yeah, but it was with a very partial bot 2010-11-22T00:34:19 and against like you and bocsimacko 2010-11-22T00:34:22 so it got pantsed 2010-11-22T00:34:34 it wasn't even expanding :P 2010-11-22T00:34:41 not that that mattered on some maps 2010-11-22T00:34:46 oh, yeah, one planet at a time bot 2010-11-22T00:34:54 hehe 2010-11-22T00:35:12 and it was really bad at going for planets on the edges of the map :| 2010-11-22T00:35:29 i've checked and rechecked the ordering or planets 2010-11-22T00:35:34 of* 2010-11-22T00:42:57 *** Azrathud has quit IRC (Ping timeout: 245 seconds) 2010-11-22T00:44:56 *** Azrathud has joined #aichallenge 2010-11-22T00:46:29 I might actually go 200 turns against mega1... 2010-11-22T00:46:42 That's a victory in my book. 2010-11-22T00:51:29 haha, i'd call that a victory 2010-11-22T00:51:53 *** McLeopold_ has quit IRC (Ping timeout: 272 seconds) 2010-11-22T00:52:52 *** Azrathud has quit IRC (Remote host closed the connection) 2010-11-22T00:57:27 *** Azrathud has joined #aichallenge 2010-11-22T01:10:08 *** McLeopold has joined #aichallenge 2010-11-22T01:10:39 okay, I would have gone 200 if my laptop didn't freeze. :( 2010-11-22T01:14:32 hehe 2010-11-22T01:15:14 i'm getting closer to having my bot made 2010-11-22T01:15:21 so i should have a few days left to tune things 2010-11-22T01:18:14 i think my response to my ranking will be along the lines of 2-20 stoked, 21-30 happy 50+ metaphorically cry like a little baby 2010-11-22T01:18:31 metaphorically? 2010-11-22T01:18:38 shhh :P 2010-11-22T01:18:39 Top 10 players: bocsimacko(3365), dmj111(3182), asavis(3172), Raschi(3150), _Astek_(3119), felixcoto(3118), davidjliu(3110), bix0r4ever(3087), _iouri_(3064), ChrisH(3058) 2010-11-22T01:19:20 it's ok.. 2010-11-22T01:19:25 women like men who cry 2010-11-22T01:20:30 no, they like men who say they are okay with crying, not actual crying men 2010-11-22T01:21:16 close enough :) 2010-11-22T01:21:20 :) 2010-11-22T01:21:44 McLeopold: i take it you're going to just wait to submit in the last few days too? 2010-11-22T01:22:05 i'm doing it at least a day before in case 2010-11-22T01:22:08 no, I'm desparatly trying to fix this pile of crap as soon as I can 2010-11-22T01:22:20 I'd like to submit now 2010-11-22T01:22:35 what's the rush? 2010-11-22T01:22:47 I want my life back. 2010-11-22T01:22:58 you can have that in a bit over a week 2010-11-22T01:23:15 Plus, I want to write a new bot. 8| 2010-11-22T01:24:25 i'm keen to try out that chess like game after this 2010-11-22T01:24:32 after the revisions for this anyway 2010-11-22T01:24:37 which one? 2010-11-22T01:24:48 umm i forget the name, it starts with a 2010-11-22T01:24:56 artimaa 2010-11-22T01:27:13 arimaa** 2010-11-22T01:27:24 almost got it :-) 2010-11-22T01:27:24 McLeopold: http://en.wikipedia.org/wiki/Arimaa that's it 2010-11-22T01:27:44 mega1: have you looked at it at all? 2010-11-22T01:27:48 os: if you want your life back.. 2010-11-22T01:27:54 you should be disqualified 2010-11-22T01:28:02 programmers arne't allowed to have lives :P 2010-11-22T01:28:15 i'm not really a programmer :P 2010-11-22T01:28:21 it's not only me who wants my life by now 2010-11-22T01:28:48 see? i knew it :P 2010-11-22T01:28:55 *** McLeopold_ has joined #aichallenge 2010-11-22T01:29:04 *** McLeopold has quit IRC (Ping timeout: 245 seconds) 2010-11-22T01:29:21 mega1: will you take the time to make a couple of revisions of your bot? mainly to handle huge numbers of planets/distances and another for n players? 2010-11-22T01:29:41 i don't expect to spend long on it, but it'd be cool to see some games of the sort 2010-11-22T01:29:49 n players is a rewrite 2010-11-22T01:30:04 hmm, i think i could get it working semi intelligably in a day or so 2010-11-22T01:30:13 one could spend ages on it of course 2010-11-22T01:30:15 huge number of planets only needs some optimization for the O(N^2) parts 2010-11-22T01:30:18 I suspect n players will shut down a lot of bots into turtle mode. 2010-11-22T01:30:34 n players is a TOTALLY different problem than 2players 2010-11-22T01:30:42 burny: yes, it's just for fun 2010-11-22T01:30:53 but it's like.. 2010-11-22T01:30:58 now that you have your checkers AI.. 2010-11-22T01:31:06 huge distances? hmm, that's easier maybe 2010-11-22T01:31:07 change a couple lines, and enter it into this chess competition 2010-11-22T01:31:34 mega1: depends, i sort of have states of noPlanets*(maxDist+2) 2010-11-22T01:31:45 so either of those numbers blowing out hurts me 2010-11-22T01:31:50 what was that game name again? 2010-11-22T01:31:54 atrima? 2010-11-22T01:31:57 arimaa 2010-11-22T01:32:29 janzert said he's planning to get back to work on it once this is over 2010-11-22T01:32:47 one would need to do a lot of training trying out different starting positions i think 2010-11-22T01:35:08 I think pong would be fun. 2010-11-22T01:35:35 it'd be hard to watch and have it not be trivial 2010-11-22T01:35:47 i guess you can slow it right down 2010-11-22T01:36:14 well, basic pong might not be fun, but it is pretty easy to add faucets to the game 2010-11-22T01:36:21 *** narnach has joined #aichallenge 2010-11-22T01:36:26 multiple balls, walls in the middle 2010-11-22T01:36:31 random maps with obstables yeah 2010-11-22T01:36:31 2d 2010-11-22T01:36:38 3d pong is awesome 2010-11-22T01:36:47 you have played that right? 2010-11-22T01:36:56 I think the only hard part would be making sure different languages can calculate the same. 2010-11-22T01:37:04 rounding errors might be an issue 2010-11-22T01:37:10 maybe 2010-11-22T01:37:16 got a link? 2010-11-22T01:37:28 http://www.newsandentertainment.com/zfcurveball.html 2010-11-22T01:37:38 I played a dos one 15 years ago 2010-11-22T01:37:56 yep, that's the one 2010-11-22T01:38:37 if you had an obstacle half way, you could hit it and keep some balls on your side, then try to overwhelm the opponenet. 2010-11-22T01:38:55 try and place one in each corner at the same time 2010-11-22T01:39:31 galcon hardest ai is real easy on almost all maps 2010-11-22T01:39:44 even though it cheats 2010-11-22T01:40:16 damn, i lost on levl 7e 2010-11-22T01:40:24 one of the better flash games ever made 2010-11-22T01:40:30 level 4: 9910 2010-11-22T01:42:17 level 6 17940 that time 2010-11-22T01:42:34 i have played that quite a bit before though 2010-11-22T01:44:38 you can ace your way through most of the first 3 levels if you put enough spin on it 2010-11-22T01:47:29 yeah, level 8: 25475 2010-11-22T01:47:33 it gets really hard by then 2010-11-22T01:47:53 15k lvl6 2010-11-22T01:47:59 first try 2010-11-22T01:49:46 it crashes on me after every game 2010-11-22T01:50:49 right click "rewind" 2010-11-22T01:52:15 rewind? 2010-11-22T01:52:39 ah, nm 2010-11-22T01:53:26 bah i give up, i got to level 8 with 4 lives left and didn;t get it once 2010-11-22T01:55:33 *** Azrathud has quit IRC (Quit: Leaving) 2010-11-22T01:56:42 bah.. 2010-11-22T01:56:46 trying to ace it sucks 2010-11-22T01:56:57 after lvl5, going for aces screws me up too often 2010-11-22T01:57:46 you need to do that in order to get the upper hand on them though 2010-11-22T02:01:07 how oes makign it real easy for them to give you an F'd up return give you the upper hand? 2010-11-22T02:01:33 not sure it's even possible to ace after lvl5 2010-11-22T02:02:27 *** amstan has quit IRC (Quit: Konversation terminated!) 2010-11-22T02:02:34 *** amstan has joined #aichallenge 2010-11-22T02:02:34 *** ChanServ sets mode: +o amstan 2010-11-22T02:10:29 it's easier to ace lvl7, tahn 6 2010-11-22T02:10:34 atleast for me 2010-11-22T02:12:12 *** bhasker_ has quit IRC (Quit: bhasker_) 2010-11-22T02:17:58 yeah, can ace past 5 2010-11-22T02:18:06 and it helps because they can't return them as well 2010-11-22T02:18:14 it might take a few moves, but it's still quicker to beat them 2010-11-22T02:18:40 Top 10 players: bocsimacko(3366), dmj111(3181), Raschi(3154), davidjliu(3145), felixcoto(3134), asavis(3125), _Astek_(3106), ChrisH(3076), _iouri_(3064), george(3058) 2010-11-22T02:20:17 I duno.. I seem to lose least often, when the serve is almsot perfectly straight 2010-11-22T02:26:26 oi.. I keep losing on lvl1 now, cause of stupid curve 2010-11-22T02:30:39 *** smloh1 has joined #aichallenge 2010-11-22T02:30:54 oh my... 2010-11-22T02:34:57 *** virdo has quit IRC (Ping timeout: 240 seconds) 2010-11-22T02:35:28 smloh1: ? 2010-11-22T02:35:33 competition fierce? 2010-11-22T02:35:52 just had my only victory against bocsi on tcp server 2010-11-22T02:36:14 joy! 2010-11-22T02:36:26 yaay! 2010-11-22T02:37:16 that's some shot of confidence, so i know i'm probably heading in the right direction 2010-11-22T02:39:36 Error323 : Your v19.0 send snippable expands at the first turn^^ 2010-11-22T02:40:42 *** Mistmanov has joined #aichallenge 2010-11-22T02:42:16 is sniping it a good idea though? or does it mean the opponent is gona lose? 2010-11-22T02:44:56 when you're snipped, you're almost sure that you're gonna loose 2010-11-22T02:49:28 here's an interesting little chart people might find interesting 2010-11-22T02:49:29 http://dl.dropbox.com/u/9140666/elo-draw-chart.png 2010-11-22T02:49:50 it is elo and draw percentage by rank 2010-11-22T02:51:00 piles of crappy bots draw 30% of the time? 2010-11-22T02:51:05 how? both crash on frame1? 2010-11-22T02:51:25 they are almost certainly pretty much all the starter package exactly 2010-11-22T02:51:30 janzert: that's awesome, but, you need a rank scale 2010-11-22T02:51:35 I think the crashers are the high peak that goes to 90% 2010-11-22T02:51:49 doesn't the starter almost never draw? 2010-11-22T02:51:54 well, it would be nice if I had one, maybe not need 2010-11-22T02:51:56 McLeopold_: ? it's the full list 2010-11-22T02:52:03 yeah, starter is assymetrical, so shouldn't draw I think 2010-11-22T02:52:09 yeah, but what rank is the 2000 elo line at? 2010-11-22T02:52:22 Mistmanov: yeah, that draw peak at the end are the crashers 2010-11-22T02:52:29 2000 elo is around 1200rank 2010-11-22T02:52:43 2200elo is 900ish rank 2010-11-22T02:53:06 1584rank is rage bot? 2010-11-22T02:53:09 actually, I think percent on the right would be percent of participants? 2010-11-22T02:53:11 the bottom has it but could certainly be done better, this is just pretty much the default chart from openoffice calc 2010-11-22T02:53:29 McLeopold_: no, right y axis is draw percentage 2010-11-22T02:53:40 oh 2010-11-22T02:53:47 left y axis is elo 2010-11-22T02:53:52 bottom is? 2010-11-22T02:53:56 rank 2010-11-22T02:54:04 got it 2010-11-22T02:54:43 whatever is rank 2900-4100.. if you remove those, the elo has a nice curve 2010-11-22T02:55:18 did that graph change with the new maps? 2010-11-22T02:55:32 I wonder how many current submissions broke? 2010-11-22T02:55:39 burny: yeah, the all the people that just submit a starter package and nothing else leave quite a plateau 2010-11-22T02:55:59 I don't think the starter has a 25% tie ratio 2010-11-22T02:56:09 McLeopold_: no, idea this is the first time I've generated it 2010-11-22T02:56:10 of the 20 or so games I saw of mine, none were a tie 2010-11-22T02:56:15 Rage bot should always draw RageBog? 2010-11-22T02:56:17 bot 2010-11-22T02:56:32 burny: only when playing against other starter packages pretty much 2010-11-22T02:56:51 McLeopold_: I think that is the peak around 1500 maybe? 2010-11-22T02:56:59 I think the starter is rank 2200-2400 2010-11-22T02:57:12 or.. do different language starter's have different logic? 2010-11-22T02:57:12 ragebo is 1500? o_O 2010-11-22T02:57:24 burny: go look at the 3000-4000 games they sure look like starter package games 2010-11-22T02:57:27 1584ish is my guess on ragebot 2010-11-22T02:57:46 at least the couple I looked at 2010-11-22T02:57:47 I think the small spike at the top is significant 2010-11-22T02:58:05 It would be cool to see a 200 game percentage too. 2010-11-22T02:58:05 I agree leo.. 2010-11-22T02:58:13 yeah it is rather interesting 2010-11-22T02:58:22 I think that the more the bots get to be 'good'.. 2010-11-22T02:58:31 the more they'll tie on all maps, always 2010-11-22T02:58:44 Perhaps it is maps that have obvious build fast a wait strategies. 2010-11-22T02:58:59 burny: depends on how many rock-paper-scissors scenarios exist on them 2010-11-22T02:58:59 http://tcp.zeroviz.us/canvas.cgi?game_id=19966 2010-11-22T02:59:16 assuming I built fast. :) 2010-11-22T02:59:24 :) 2010-11-22T02:59:54 *** Palmik has joined #aichallenge 2010-11-22T02:59:58 But, that type of map is likely to draw a lot against good bots. 2010-11-22T03:00:11 (And should totally be kept in the contest.) 2010-11-22T03:00:41 ya, as long as there isn't lots of that type 2010-11-22T03:00:43 *** yasith_ has quit IRC (Ping timeout: 245 seconds) 2010-11-22T03:00:44 should be a couple 2010-11-22T03:00:48 *** Rubicon-|-Cross has joined #aichallenge 2010-11-22T03:00:48 So, my guess would be 2-3% of maps are obvious strategies. 2010-11-22T03:00:56 *** Rubicon-|-Cross has quit IRC (Read error: Connection reset by peer) 2010-11-22T03:01:18 among top 100 vs 100 games the highest draw rate is still only 6.9 percent with map_r538_39 2010-11-22T03:01:20 the elo line is interesting 2010-11-22T03:01:24 *** Rubicon-|-Cross has joined #aichallenge 2010-11-22T03:01:32 *** Rubicon-|-Cross has quit IRC (Client Quit) 2010-11-22T03:01:40 *** Rubicon-|-Cross has joined #aichallenge 2010-11-22T03:01:44 how about top10? 2010-11-22T03:01:53 does that flat spot make it tough to move up if your in there? 2010-11-22T03:01:57 *** narnach has quit IRC (Quit: Leaving.) 2010-11-22T03:02:04 since top100 has a pretty huge range of goodness 2010-11-22T03:02:21 leo, 'tough' no.. just slow 2010-11-22T03:02:26 takes like 10 games to get through it 2010-11-22T03:02:32 http://tcp.zeroviz.us/map_r538_39.txt.html 2010-11-22T03:02:37 burny: no idea and it wouldn't be that many games to draw much of a conclusion from anyway 2010-11-22T03:03:37 same map: http://tcp.zeroviz.us/canvas.cgi?game_id=20089 2010-11-22T03:04:40 I'm almost thinking that instead of taking the effort to do a proper bot.. 2010-11-22T03:04:41 even though it draws a lot, there is plenty of room for strategy in that map 2010-11-22T03:04:52 I want to just tweak, and see how high I can get a gimpy simple bot 2010-11-22T03:05:03 another one that ties it is map_r538_44 2010-11-22T03:05:35 not on my server 2010-11-22T03:05:50 and there are a couple more that I'm sure don't have a statisically significant difference 2010-11-22T03:06:06 this map http://ai-contest.com/visualizer.php?game_id=8292400 2010-11-22T03:06:17 I wonder if I can break the top 100, with a gimpy bot 2010-11-22T03:06:21 ahh, very close start 2010-11-22T03:06:55 that 4 in the middle is just waiting to get sniped 2010-11-22T03:06:58 although that central 4 neutral may make it an r-p-s starting situation? 2010-11-22T03:07:18 probably not possible to get into top100, without making the bot consider what the opponent is doing(ie. defend against attacks) 2010-11-22T03:07:32 burny: yeah, probably not anymore 2010-11-22T03:07:40 defending against attacks doesn't really predict the opponent 2010-11-22T03:07:56 it's more of a reaction 2010-11-22T03:07:57 leo, if I have 13 ships at a 5grow planet.. 2010-11-22T03:08:04 and the opponent sends 65ships at it.. 2010-11-22T03:08:21 my current bot waits until it's captured 10turns later, and then sends 65ships at it 2010-11-22T03:08:25 McLeopold_: I think he simply means predicting future states 2010-11-22T03:08:40 oh 2010-11-22T03:09:43 and if after sending those 65.. the opponent sends 5.. I don't care about those 5, until 10turns later 2010-11-22T03:09:44 etc. 2010-11-22T03:10:35 yeah, I don't think you'll go very high without future planet states 2010-11-22T03:10:52 at 2100ish elo 2010-11-22T03:11:14 can get to 2500ish I'm sure 2010-11-22T03:11:23 but higher then that, I duno 2010-11-22T03:11:50 it's not very hard to implement future states 2010-11-22T03:12:15 it is compared to what I have 2010-11-22T03:12:56 :P 2010-11-22T03:13:27 My first bot didn't have future states. 2010-11-22T03:13:51 *** yasith_ has joined #aichallenge 2010-11-22T03:14:04 i think it actually makes it simpler to string together complex logic once you implement it 2010-11-22T03:14:15 your bot can think at a higher level 2010-11-22T03:14:35 my bot consists of like.. 2010-11-22T03:14:54 *** medrimonia1 has quit IRC (Read error: Connection reset by peer) 2010-11-22T03:15:08 if(bestTargetPlanet().population < myPlanet.population) attack 2010-11-22T03:15:24 inside a do while 2010-11-22T03:15:48 burny: put your bot up on tcp2.zeroviz.us 2010-11-22T03:15:56 well soon you'll be adding on "+growthrate*distance" 2010-11-22T03:16:06 I'll but my first bot on there and we can see how they match up 2010-11-22T03:16:12 and before you know it you'll have a mess of calculations 2010-11-22T03:16:35 *** medrimonia1 has joined #aichallenge 2010-11-22T03:16:37 growrate is handled by 'bestTargetPlanet' 2010-11-22T03:17:03 ha i did, but i think i timed out 2010-11-22T03:17:46 bah, that's not good 2010-11-22T03:18:08 leo, maybe tomorrow 2010-11-22T03:18:10 need sleep 2010-11-22T03:18:39 hmm, it doesn't run 2010-11-22T03:18:41 Top 10 players: bix0r4ever(3411), bocsimacko(3402), davidjliu(3267), dmj111(3200), Raschi(3193), medrimonia(3191), asavis(3169), ChrisH(3122), felixcoto(3121), _Astek_(3120) 2010-11-22T03:18:45 yeah, later 2010-11-22T03:19:09 *** tapwater has quit IRC (Quit: tapwater) 2010-11-22T03:19:24 bix wans't even on top10 a few mins ago? 2010-11-22T03:19:42 i just reduced the time limit 2010-11-22T03:20:02 he has been top10 for awhile 2010-11-22T03:20:40 bah, it died again 2010-11-22T03:20:55 i was beating mcleopold too aha 2010-11-22T03:20:56 your 11 turns looked promising. :) 2010-11-22T03:21:11 it sin't even redistributing or making planets safe yet :P 2010-11-22T03:22:13 burny: bix just resubmitted a little while ago, so wouldn't have been in the rankings for a little bit 2010-11-22T03:22:15 man, to think that bot was #1 for a week... 2010-11-22T03:22:43 ah god damnit 2010-11-22T03:22:50 my bot is terrible, i wonder what's causing that 2010-11-22T03:22:55 i hope it's not invalid moves 2010-11-22T03:23:05 you are playing on the crazy maps 2010-11-22T03:23:25 it shouldn't crash 2010-11-22T03:23:49 you pantsed me then 2010-11-22T03:24:01 it is invalid orders 2010-11-22T03:24:09 hmm, i was just messing around trying to get redistribution going 2010-11-22T03:24:14 i must have screwed something up 2010-11-22T03:24:17 hopefully it is just from that 2010-11-22T03:25:00 yeah, if map is to go forward, the old bot rocked 2010-11-22T03:25:03 right, i'm going to work out what's causing that 2010-11-22T03:27:22 bah, that time it crashed :| 2010-11-22T03:27:25 this isn't good 2010-11-22T03:29:01 McLeopold_: it seems to do alright against you until crashing 2010-11-22T03:32:31 that's good 2010-11-22T03:33:56 what's the maximum number of planets for these maps? 2010-11-22T03:34:03 the one we are playing on now has quite a fet 2010-11-22T03:34:05 few* 2010-11-22T03:36:19 40 2010-11-22T03:36:32 41 2010-11-22T03:36:45 http://tcp2.zeroviz.us/canvas.cgi?game_id=1072 2010-11-22T03:36:47 big ass map 2010-11-22T03:37:18 yeah, but not much growth 2010-11-22T03:38:10 bah it crashed again 2010-11-22T03:40:17 it needs redistribution desperately 2010-11-22T03:40:48 you should be able to just drop in a funneling system 2010-11-22T03:41:00 i already have the strategy worked out 2010-11-22T03:41:03 i'm just not rushing things :) 2010-11-22T03:43:22 is that your really old submit bot? 2010-11-22T03:46:38 it's making some silly moves 2010-11-22T03:55:13 I got a math question 2010-11-22T03:55:26 ... 2010-11-22T03:55:28 let's say you have a growth 4 planet with 4 ships right in the middle 2010-11-22T03:55:33 what do you do? 2010-11-22T03:55:56 it depends how close you are to them 2010-11-22T03:56:01 and what other options are available 2010-11-22T03:56:04 if you both go for it, and the enemy sends more, you lose 4 ships for nothing 2010-11-22T03:56:15 assume it's the best thing going 2010-11-22T03:56:30 send everything that doesn't leave your other planets vulnerable 2010-11-22T03:56:32 distance 6 2010-11-22T03:57:07 although, it could be the case that then you lose if they send everything and you don't have enough left to claim their other planets 2010-11-22T03:57:08 okay, assume there are a few other things out there, send everything that is left? 2010-11-22T03:57:13 it really can change depending on the situation 2010-11-22T03:58:50 by bot is having a hard time figuring this out 2010-11-22T03:58:51 *** Ice_Harley has joined #aichallenge 2010-11-22T03:59:00 which means I am.. 2010-11-22T03:59:02 i think my bot would consider sending everything it can at is 2010-11-22T03:59:04 at it* 2010-11-22T03:59:10 but also any other moves and doing nothing 2010-11-22T03:59:12 @ranking 2010-11-22T03:59:13 Ice_Harley: Top 10 players: bocsimacko(3369), bix0r4ever(3248), davidjliu(3225), dmj111(3201), Raschi(3188), asavis(3178), felixcoto(3122), Ice_Harley(3112), medrimonia(3099), oldman(3089) 2010-11-22T03:59:16 then let the game tree pick what works best 2010-11-22T03:59:38 I don't do game trees... how did you get it to work? 2010-11-22T03:59:57 You have to assume the opponent moves like you, right? 2010-11-22T04:00:21 yeah, although i don't consider them expanding 2010-11-22T04:00:45 at each node, i pick moves for the player, also adding the zero move, then make children for each of those 2010-11-22T04:01:13 generate the tree for 0.9 seconds, then evaluate the terminal nodes and use minimax to pick the best string of strategies 2010-11-22T04:01:17 then implement those 2010-11-22T04:01:18 I could probably do with some prediction 2010-11-22T04:01:39 I could stick the moves in, and my framework handles it, I've just never tried it. 2010-11-22T04:02:00 It might help with first turn decisions 2010-11-22T04:02:18 yeah, but it's hard to have them expand 2010-11-22T04:02:19 But, so far, it seems a little better to just react. 2010-11-22T04:02:23 but not try to snipe their "mock moves" 2010-11-22T04:02:28 Ice_Harley: Gratz 2010-11-22T04:02:31 didn't think of that 2010-11-22T04:02:34 but also snipe their actual moves 2010-11-22T04:02:54 so i just don't have them expanding at all, only defending/attacking 2010-11-22T04:02:59 janzert: so much for neutrality :-) 2010-11-22T04:03:53 heh, gotta give out something for getting into the top 10 ;) 2010-11-22T04:04:11 mega1: you wanna chime in on the unsnipable planet delima? 2010-11-22T04:04:28 sure 2010-11-22T04:04:36 where is it? 2010-11-22T04:04:48 centre planet, 4 growth and 4 ships 2010-11-22T04:05:11 http://tcp.zeroviz.us/canvas.cgi?game_id=19946 2010-11-22T04:05:16 that's not general enough for an exact strategy, but dependent on your own given situations with that, what do you do? 2010-11-22T04:05:33 mega1: baits the opponent... 2010-11-22T04:05:41 *** mceier has joined #aichallenge 2010-11-22T04:06:22 i'm confused 2010-11-22T04:06:31 *** smloh1 has quit IRC (Quit: Page closed) 2010-11-22T04:06:31 mega1 also doesn't worry about his home planet safty 2010-11-22T04:06:31 how did you decide to do that third turn and not second? 2010-11-22T04:06:44 I know 2010-11-22T04:06:50 nevermind, fleet is invisible 2010-11-22T04:06:51 turn 1, wait for dominance in the area 2010-11-22T04:07:02 turn 2, send enough to take the planet 2010-11-22T04:07:06 he sends 5 second turn, that was dangerous ? 2010-11-22T04:07:12 I'm still trying to digest it 2010-11-22T04:07:14 what if error send 6? 2010-11-22T04:07:17 turn 3, assume enemy will send as well and jump? 2010-11-22T04:07:32 what happens if error sent 6 there on the second turn 2010-11-22T04:07:49 his home planet might then be vulnerable, i think he'd still be snipable? 2010-11-22T04:08:06 http://tcp.zeroviz.us/canvas.cgi?game_id=12157 2010-11-22T04:08:30 without dominance, mega1 just goes for it turn 2? 2010-11-22T04:08:46 oh, wait, he sniped 2010-11-22T04:09:00 and switched 2010-11-22T04:09:09 yeah, you went for it 2010-11-22T04:09:24 the optimal strategy there is to wait i'm pretty sure 2010-11-22T04:09:26 in the first there is nothing that can be taken over without losing that 4 2010-11-22T04:09:55 what about the 98? not important to you :) 2010-11-22T04:10:24 wait, if you send 5 on the first turn and they wait, isn't it kind of rps? 2010-11-22T04:10:42 nope, look what happens 2010-11-22T04:10:54 not if you back up your claim and they try to snipe 2010-11-22T04:11:05 I get back to even next turn, but I lose 4 growth after that every turn 2010-11-22T04:11:12 Even with trying to go for the home planet 2010-11-22T04:11:23 you send nothing to 4 neutral on the second turn 2010-11-22T04:11:31 you could have saved his snipe 2010-11-22T04:11:39 that's the point of it being ships <= growth 2010-11-22T04:11:54 but he might have predicted you would do that and send at your home planet 2010-11-22T04:12:03 yeah, but how do you know to do that generically? 2010-11-22T04:12:09 on this map, sure 2010-11-22T04:12:37 i think you can take that 4 then have a very undecided match on your hands 2010-11-22T04:13:00 mega1: what do you think? 2010-11-22T04:13:23 I think so too 2010-11-22T04:14:05 McLeopold_: the answer is there is no immediate way to handle your given condition, it's too general 2010-11-22T04:14:33 *** perror has joined #aichallenge 2010-11-22T04:16:00 it's the simple 2 player 2 planet isosceles triangle example, which is far from trivial in itself 2010-11-22T04:16:05 3 planet* 2010-11-22T04:16:29 that is another type of map i would like to do, randomly made 3 planet maps 2010-11-22T04:16:49 it seems like jumping is the way to go 2010-11-22T04:17:29 if you placed all planets equally spaced around a circle, the game would look cool 2010-11-22T04:17:49 especially if fleets started flying everywhere 2010-11-22T04:18:40 Top 10 players: bocsimacko(3371), bix0r4ever(3315), davidjliu(3274), _Astek_(3239), dmj111(3211), medrimonia(3184), Raschi(3183), asavis(3175), Ice_Harley(3169), ChrisH(3160) 2010-11-22T04:31:29 *** phreeza has joined #aichallenge 2010-11-22T04:32:10 *** narnach has joined #aichallenge 2010-11-22T04:32:10 *** narnach is now known as narnach_mb 2010-11-22T04:44:59 *** medrimonia has quit IRC (Ping timeout: 264 seconds) 2010-11-22T04:50:55 *** McLeopold_ has quit IRC (Ping timeout: 245 seconds) 2010-11-22T04:51:15 *** phreeza has quit IRC (Quit: Computer has gone to sleep) 2010-11-22T04:53:44 *** Naktibalda has joined #aichallenge 2010-11-22T04:57:00 *** medrimonia has joined #aichallenge 2010-11-22T05:12:25 bah, i have a bug somewhere 2010-11-22T05:12:29 and i'm not sure where 2010-11-22T05:18:15 *** mega1 has quit IRC (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.) 2010-11-22T05:18:41 Top 10 players: bocsimacko(3370), Raschi(3287), medrimonia(3269), dmj111(3258), bix0r4ever(3189), felixcoto(3179), ChrisH(3173), davidjliu(3167), asavis(3166), Ice_Harley(3165) 2010-11-22T05:30:45 *** Rubicon-|-Cross has quit IRC (Quit: Leaving) 2010-11-22T05:42:07 bah, i get invalid orders when i'm redistributing 2010-11-22T05:45:37 *** phreeza has joined #aichallenge 2010-11-22T05:57:43 *** phreeza has quit IRC (Quit: kthxbai) 2010-11-22T06:02:45 *** Mistmanov has quit IRC (Ping timeout: 265 seconds) 2010-11-22T06:18:42 Top 10 players: Raschi(3338), bocsimacko(3338), _Astek_(3245), dmj111(3224), medrimonia(3213), asavis(3194), felixcoto(3185), davidjliu(3175), Manwe56(3130), wagstaff(3129) 2010-11-22T06:58:51 it's been a while since bosimacko was third 2010-11-22T06:58:58 even if just from other people having new submissions 2010-11-22T07:02:35 sweet, my deadline is now 11:59pm saturday 2010-11-22T07:02:45 it closes at like 4pm sunday for my time 2010-11-22T07:18:42 Top 10 players: Raschi(3480), _Astek_(3364), bocsimacko(3356), dmj111(3240), asavis(3197), felixcoto(3184), medrimonia(3173), davidjliu(3164), BaronTrozo(3162), Manwe56(3122) 2010-11-22T07:19:34 *** antimatroid1 has joined #aichallenge 2010-11-22T07:21:33 *** antimatroid has quit IRC (Ping timeout: 245 seconds) 2010-11-22T07:42:35 say me please the best target selection heuristic 2010-11-22T07:43:12 hahah 2010-11-22T07:43:21 you tell me ;) 2010-11-22T07:44:24 my target selection heuristic something strange irracinal and bad working now 2010-11-22T07:45:00 my target selection heuristic something strange irraci0nal and bad working now 2010-11-22T07:45:04 my target selection heuristic something strange irracional and bad working now 2010-11-22T07:46:41 *** phreeza has joined #aichallenge 2010-11-22T07:47:20 Accoun: i just scan through the planets, go through all time states and make moves for the planet 2010-11-22T07:52:46 antimatroid1> what the best planet to make moves? 2010-11-22T07:53:02 antimatroid1> how to seleck best target? 2010-11-22T07:53:08 antimatroid1> how to select best target? 2010-11-22T07:53:09 i sort planets firstly by growth then secondly by average distance to other planets 2010-11-22T07:54:06 *** Appleman1234 has quit IRC (Quit: Leaving) 2010-11-22T07:56:09 antimatroid1> number of neutral ships on it is effect 2010-11-22T07:56:13 antimatroid1> number of neutral ships on it is effect? 2010-11-22T07:56:23 antimatroid1> number of neutral ships on it is any effect? 2010-11-22T07:56:25 i take that into account when i make moves and assign them a payoff 2010-11-22T07:56:34 although atm, i'm considering every possible move at each node 2010-11-22T07:57:17 antimatroid1> effect= proffit/payoff ? 2010-11-22T07:58:35 sorry waht? 2010-11-22T07:59:16 antimatroid1> effect= evalation/payoff ? 2010-11-22T07:59:25 i don't know what you mean 2010-11-22T08:01:19 i hate when someone floods while trying to talk 2010-11-22T08:04:40 some bots have realy good traget selection heuristic or use some kaind of alfa-beta algoritm 2010-11-22T08:04:50 jup 2010-11-22T08:04:55 Accoun: are you just using minimax? 2010-11-22T08:05:14 i'm playing as kirby atm, but it's still not "making planets safe" before making moves 2010-11-22T08:05:21 and it's got some bug making invalid orders :\ 2010-11-22T08:11:40 i think there's a lot of submissions happening now, so people aren't getting many games 2010-11-22T08:11:48 i suspect it'll be like that for the rest of the week 2010-11-22T08:18:44 Top 10 players: Raschi(3527), _Astek_(3402), bocsimacko(3375), dmj111(3256), BaronTrozo(3222), asavis(3212), felixcoto(3193), medrimonia(3169), davidjliu(3165), Manwe56(3125) 2010-11-22T08:19:04 antimatroid1> thanks for ideas 2010-11-22T08:20:22 Accoun: what do you use currently then? 2010-11-22T08:21:18 I have depression surge, cant work cant play cant sleep 2010-11-22T08:21:37 uhhh? 2010-11-22T08:21:48 Accoun: are you just using minimax? 2010-11-22T08:21:52 i don't trust alpha beta 2010-11-22T08:21:57 i may try it before the weekend though 2010-11-22T08:22:04 antimatroid1: how can you not trust alphabeta :P 2010-11-22T08:22:19 its not really about trust 2010-11-22T08:22:20 hehe 2010-11-22T08:22:27 what do you mean 2010-11-22T08:22:36 i trust it when evaluating a tree already made, but that's almost free for me, i don't trust cropping my tree based off incremental payoffs 2010-11-22T08:22:47 i might just implement it to cut stuff off for terminal states 2010-11-22T08:23:00 as i generate it 2010-11-22T08:23:09 *** delt0r has joined #aichallenge 2010-11-22T08:23:43 *** delt0r___ has quit IRC (Ping timeout: 265 seconds) 2010-11-22T08:23:57 although once i find a terminal state that's already made a terminal node, but i can probably get some information about that for the layer above 2010-11-22T08:24:58 yeah, i think i shall implement alpha beta, but only with pruning where terminal states have been reached, if that makes sense at all 2010-11-22T08:25:22 i pimped my neutral capping algo :D 2010-11-22T08:25:27 i still need to add some things before i can worry about things like that though 2010-11-22T08:25:35 my bot is only partially complete 2010-11-22T08:26:18 it got wins against testingFromPiter and manwe56 before, but threw some games pretty badly 2010-11-22T08:26:22 antimatroid1: don't you worry about the deadline? 2010-11-22T08:26:28 *** mceier has quit IRC (Quit: leaving) 2010-11-22T08:26:36 Error323: i have till sunday morning, and nothing to do till then 2010-11-22T08:26:40 it'll be done by then 2010-11-22T08:26:49 sunday morning my time anyway 2010-11-22T08:27:04 i will probably have my final one uploadeded before bed saturday night 2010-11-22T08:29:20 antimatroid1> wat is name of u bot? 2010-11-22T08:29:34 youwillfail 2010-11-22T08:29:40 i don't have anything uploaded to the main site atm 2010-11-22T08:30:06 antimatroid1> playing on TCP ? 2010-11-22T08:30:13 kirby 2010-11-22T08:30:27 but it still needs a few more things added in before i'd call it finished and ready for final tunings 2010-11-22T08:30:37 Accoun: are you running linux? 2010-11-22T08:30:48 and i need to work out what is causing invalid orders :\ 2010-11-22T08:31:40 antimatroid1> testingFromPiter2 also my bot 2010-11-22T08:31:57 Accoun: if you are running linux, wanna trade binaries? 2010-11-22T08:31:58 you are testingFromPiter? 2010-11-22T08:32:03 antimatroid1> runing on netbook without logs save 2010-11-22T08:32:18 that solves that unsolved problem? 2010-11-22T08:32:22 antimatroid1> it ols varsion for comparation 2010-11-22T08:32:27 antimatroid1> it old varsion for comparation 2010-11-22T08:33:05 Accoun: the leaderboard is going to be live during the tournament this time :) 2010-11-22T08:33:33 antimatroid1> only 50 maps 2010-11-22T08:33:47 Accoun: hellooo? 2010-11-22T08:33:48 =) 2010-11-22T08:34:18 i don't get to decide that, but there will be a decent selection 2010-11-22T08:34:22 http://72.44.46.68/canvas?game_id=1141932 :P from before 2010-11-22T08:34:28 Error323> windows XP 2010-11-22T08:34:52 wtfffffff guys what's with this windows love in this channel >_<< 2010-11-22T08:34:55 antimatroid1> good heuristic 2010-11-22T08:35:05 * antimatroid1 is using xp too 2010-11-22T08:35:07 antimatroid1> better tham my 2010-11-22T08:35:21 it still sends fleets out stupidly sometimes 2010-11-22T08:35:29 error caned me before cause of stupid moves like that 2010-11-22T08:37:23 mm? 2010-11-22T08:37:42 how do you guys pick redistribution planets? 2010-11-22T08:37:55 *** Palmik has quit IRC (Read error: Connection reset by peer) 2010-11-22T08:37:57 haha windows love 2010-11-22T08:38:07 i pick the planet closer to the enemy that will be mine when the fleet gets there that minimises the distance to it and then it's closest enemy 2010-11-22T08:38:12 it seems to do a reasonable job 2010-11-22T08:38:24 and i wont send to a plant further than the sources closest enemy 2010-11-22T08:39:26 it creates cool trees anyway :P 2010-11-22T08:39:37 nice work antimatroid1 2010-11-22T08:40:29 Error323: it's not too far away from being done 2010-11-22T08:40:33 hence the lack of worry about the deadline 2010-11-22T08:40:40 i understand 2010-11-22T08:40:55 i'm getting my hopes up for top 10 again :S 2010-11-22T08:41:00 just don't beat mine 2010-11-22T08:41:01 top 30 would be nice 2010-11-22T08:41:18 or I will visit tasmania 2010-11-22T08:41:21 ^_^ 2010-11-22T08:41:25 haha 2010-11-22T08:41:58 antimatroid1> u name on oficial? 2010-11-22T08:42:03 youwillfail 2010-11-22T08:42:12 http://ai-contest.com/profile.php?user_id=3945 2010-11-22T08:42:40 it says i was in a partnership, but my friend never ended up helping out 2010-11-22T08:42:40 i hate submitting and losing first two games :P 2010-11-22T08:42:46 i don't think i'll be able to get that changed 2010-11-22T08:42:56 antimatroid1> Nov 3rd 09:29:39 error while compiling submission C++ 2010-11-22T08:43:25 yeah, my old bot got separated into multiple folders which the server hated, then i had exams and started on a fresh bot, which is still in construction 2010-11-22T08:44:00 that's from the start of the month 2010-11-22T08:44:32 i'm surprised astek is still on top 2010-11-22T08:45:36 my boot need tuning 2010-11-22T08:45:47 aloot of tuning 2010-11-22T08:46:11 so does mine, but priority one is getting it fixed 2010-11-22T08:46:18 finished* 2010-11-22T08:46:32 forses of atomic nuclear armagedon- give me the power! 8-) 2010-11-22T08:46:40 lol 2010-11-22T08:47:23 is that the russian version of he-man? 2010-11-22T08:48:07 lmao 2010-11-22T08:48:18 and skeletor? 2010-11-22T08:48:25 i only know about that from cky 2010-11-22T08:49:32 http://www.youtube.com/watch?v=7yeA7a0uS3A 2010-11-22T08:50:09 http://www.youtube.com/watch?v=T3vYJSzOXmE 2010-11-22T08:51:26 *** Sylph2 has joined #aichallenge 2010-11-22T08:51:51 haha wtf 2010-11-22T08:52:15 back when bam margera wasn't a sold out douche 2010-11-22T08:53:25 *** Sylph has quit IRC (Ping timeout: 245 seconds) 2010-11-22T08:55:14 lazzy town the best 2010-11-22T08:55:26 *** Palmik has joined #aichallenge 2010-11-22T08:55:34 mult serial for boys not so good 2010-11-22T08:56:28 http://www.youtube.com/watch?v=FPcXkR8moBs 2010-11-22T08:57:19 *** narnach has joined #aichallenge 2010-11-22T08:57:39 *** bhasker has joined #aichallenge 2010-11-22T09:01:30 *** Utkarsh has joined #aichallenge 2010-11-22T09:07:31 Accoun: are they singing about piroggi? 2010-11-22T09:07:52 Accoun: is that russian? 2010-11-22T09:07:54 i have seen it before 2010-11-22T09:09:07 antimatroid1> russian version 2010-11-22T09:09:46 http://www.youtube.com/watch?v=wos2jhKhiCQ&NR=1 2010-11-22T09:09:48 lol what 2010-11-22T09:10:19 why on earth does that have 15M views?? 2010-11-22T09:10:59 http://www.youtube.com/watch?v=9m6Z81BQk44&feature=related 2010-11-22T09:12:19 its lazy town time! 2010-11-22T09:12:30 phreeza: it's the internet, don't ask 2010-11-22T09:13:06 i can't believe that has 5 million views 2010-11-22T09:13:19 i can't believe i'm contributing to that 2010-11-22T09:13:31 yea 2010-11-22T09:13:41 back to the bot i think 2010-11-22T09:13:50 *** p4p4 has joined #aichallenge 2010-11-22T09:18:43 Top 10 players: bocsimacko(3388), _Astek_(3275), dmj111(3244), asavis(3234), Ice_Harley(3156), luksian(3153), davidjliu(3153), felixcoto(3150), medrimonia(3149), Manwe56(3123) 2010-11-22T09:29:36 morning 2010-11-22T09:31:09 howdy 2010-11-22T09:31:42 i still can't figure out how to calculate available ships correctly so freaking annoying 2010-11-22T09:31:59 i hate it when my bot won't defend a planet from a nearby planet with lots of ships 2010-11-22T09:32:21 grrr 2010-11-22T09:32:57 that's the last major thing i need to add, then it's all tuning 2010-11-22T09:33:33 http://www.youtube.com/watch?v=3AzpByR3MvI 2010-11-22T09:33:33 how are you calculating available ships 2010-11-22T09:33:37 i have tried various models 2010-11-22T09:33:40 none of them work well 2010-11-22T09:34:15 i take it you're accounting for incoming fleets already? 2010-11-22T09:34:19 yes 2010-11-22T09:34:26 the problem is when there is nothing incoming 2010-11-22T09:34:32 what is a good number you can safely send out 2010-11-22T09:34:51 so the nearby planet will have ships, and no incoming 2010-11-22T09:34:57 then my other hack is to have them mock send their ships to the closest planets draining away their usable ships, then picking moves, then resetting the moves they mock made 2010-11-22T09:34:59 but the way i calculate available ships probably locks it down 2010-11-22T09:35:07 hmm 2010-11-22T09:35:28 i tried a big thing where i'd try to make "all planets safe" like that, but it ended up just crippling it 2010-11-22T09:35:34 yea 2010-11-22T09:35:55 if i assume everything is available unless under attack it ends up redirecting ships from front line planets 2010-11-22T09:35:56 i suspect i'll probably only mock send the fleets they have at time 0 or the first few turns too 2010-11-22T09:36:09 i tried that didnt work very well 2010-11-22T09:36:20 it used to work alright for me 2010-11-22T09:36:24 i hope it still does 2010-11-22T09:36:50 hmm 2010-11-22T09:39:55 damnit tcp is sooo damn random 2010-11-22T09:39:55 >_< 2010-11-22T09:41:42 Error323> random for fist 8 hours of testing 2010-11-22T09:42:00 yea 2010-11-22T09:46:40 every tweak takes 8 hours of testing 2010-11-22T09:46:42 yuck 2010-11-22T09:54:54 i have kind of given up now 2010-11-22T09:57:11 hmm 2010-11-22T09:57:17 well i have a few more ideas 2010-11-22T09:57:23 but ughhh the time it takes :P 2010-11-22T09:57:26 to test 2010-11-22T10:05:33 *** rebelxt has left #aichallenge 2010-11-22T10:05:49 *** JamesMG has joined #aichallenge 2010-11-22T10:06:20 JamesMG: why can't you lose consistently 2010-11-22T10:07:29 because my bot is not consistant :) 2010-11-22T10:07:58 can never protect planets! 2010-11-22T10:08:05 or be 100% safe! 2010-11-22T10:08:17 in fact it's probably a good idea to underprotect a specific planet 2010-11-22T10:08:24 attract the enemy bot to that one planet 2010-11-22T10:12:28 sun tzu bot.. 2010-11-22T10:12:41 j3camero: get a chance to play slay yet? :P 2010-11-22T10:14:22 somehow my official bot has been gaining ranks, despite being submitted on Nov 8 2010-11-22T10:14:32 err, 11th rather 2010-11-22T10:14:48 not that I'm complaining 2010-11-22T10:14:53 *** Sylph2 has quit IRC (Ping timeout: 245 seconds) 2010-11-22T10:14:57 at least it wasn't submitted oct 8.. 2010-11-22T10:15:28 so theoretically it's a polish up bot? ;) 2010-11-22T10:15:31 polished* 2010-11-22T10:17:18 *** lavalamp_ has quit IRC (Quit: Page closed) 2010-11-22T10:17:46 *** fawek has joined #aichallenge 2010-11-22T10:18:44 Top 10 players: bocsimacko(3368), Raschi(3358), dmj111(3213), asavis(3208), medrimonia(3174), _Astek_(3169), felixcoto(3139), bix0r4ever(3129), davidjliu(3127), BaronTrozo(3116) 2010-11-22T10:20:10 no, it has a very long ways to go 2010-11-22T10:20:18 I just make it worse every time I try to go with it 2010-11-22T10:21:07 I guess everyone is back on dhartmei's TCP server? 2010-11-22T10:21:13 ye 2010-11-22T10:22:46 wish I had more time this week to work on it 2010-11-22T10:22:57 my current submission might end up being my final one 2010-11-22T10:34:33 *** Utkarsh has quit IRC (Ping timeout: 272 seconds) 2010-11-22T10:40:06 *** Utkarsh has joined #aichallenge 2010-11-22T10:44:13 *** Utkarsh has quit IRC (Ping timeout: 240 seconds) 2010-11-22T10:50:11 *** Utkarsh has joined #aichallenge 2010-11-22T10:56:40 *** Utkarsh has quit IRC (Ping timeout: 240 seconds) 2010-11-22T11:01:51 *** Utkarsh has joined #aichallenge 2010-11-22T11:14:59 ha, ha, ha, ha, ha hah ahaha 2010-11-22T11:15:01 all this time 2010-11-22T11:15:17 i've been evaluating terminal nodes where they made the last move as though i had 2010-11-22T11:15:24 ahahaha 2010-11-22T11:16:35 *** Mistmanov has joined #aichallenge 2010-11-22T11:17:55 so today I finally decided to stop procrastinating, and start working on my report for school instead of on this AI contests... 2010-11-22T11:18:13 ...3 minutes into the bikeride to the university, I get this brilliant idea to improve my bot -_- 2010-11-22T11:18:36 * Mistmanov goes back to coding 2010-11-22T11:18:44 Top 10 players: Raschi(3411), bocsimacko(3374), dmj111(3220), asavis(3189), medrimonia(3173), _Astek_(3162), luksian(3140), davidjliu(3115), felixcoto(3104), wagstaff(3104) 2010-11-22T11:18:46 don' 2010-11-22T11:18:51 don't fight the brilliant ideas :P 2010-11-22T11:22:30 false alarm, my payoff function was fine 2010-11-22T11:22:35 * antimatroid1 feels stupid now 2010-11-22T11:22:48 * janzert idly wonders what the average grade point reduction is as the result of the contest :P 2010-11-22T11:23:04 *** mceier has joined #aichallenge 2010-11-22T11:23:41 jaznert: i'll let you know thursday :P 2010-11-22T11:23:52 I have the "luck" that my report has no deadline 2010-11-22T11:23:53 i don't expect my normal results 2010-11-22T11:24:05 so I can just keep procrastinating as long as I keep paying to be in school ;) 2010-11-22T11:24:10 damn, i'm good at jaznert typo 2010-11-22T11:24:15 :) 2010-11-22T11:24:17 i think you should just change your name :P 2010-11-22T11:25:52 *** Cold-Phoenix has joined #aichallenge 2010-11-22T11:28:31 *** Ice_Harley has quit IRC (Ping timeout: 265 seconds) 2010-11-22T11:30:24 * medrimonia1 is loosing something between 5 and 10 over 20 points on his grades 2010-11-22T11:33:38 *** McLeopold has joined #aichallenge 2010-11-22T11:33:43 @tcp 2010-11-22T11:33:44 McLeopold: tcp = (#1) http://www.benzedrine.cx/planetwars/, or (#2) http://72.44.46.68/, or (#3) 98.247.248.39:995 (zeroviz.us), or (#4) http://tcp.zeroviz.us/, or (#5) http://tcp2.zeroviz.us/ (crazy!) 2010-11-22T11:33:50 *** bhasker has quit IRC (Quit: bhasker) 2010-11-22T11:34:05 *** McLeopold has quit IRC (Client Quit) 2010-11-22T11:34:44 *** TEAM_AI has joined #aichallenge 2010-11-22T11:34:49 hi 2010-11-22T11:35:35 can anyone tell how to i can implement bot-logging in c++ ? 2010-11-22T11:35:46 ofstreams 2010-11-22T11:38:15 it won't cause errs? 2010-11-22T11:38:27 you can't submit it with that 2010-11-22T11:39:39 it's terrible that cout or similar doesnt work ^^ 2010-11-22T11:39:54 that's how you communicate with the engine 2010-11-22T11:41:22 just make a logger class, create an instance in main and refer to it throughout ur project with a macro 2010-11-22T11:42:08 and then i can send strings to console at PlayGame? 2010-11-22T11:42:50 or is it better to use files? 2010-11-22T11:43:12 no idea whether it's handy or anything, but I do all my logging via cerr 2010-11-22T11:43:43 and then cygwin sends the cerr thingie to file 2010-11-22T11:44:42 hm, damn ms vs ^^ is it possible to use windows msgboxes? 2010-11-22T11:47:03 and one important q: timlimit per round is 1 sec how fast are the challange machines? 2010-11-22T11:47:03 *** nann has quit IRC (Read error: Connection reset by peer) 2010-11-22T11:47:30 *** Apophis has joined #aichallenge 2010-11-22T11:49:48 *** narnach_mb has quit IRC (Quit: Leaving.) 2010-11-22T11:50:42 TEAM_AI: you can use anything you like on your machinem just put sime #ifdefs around the relevant code... 2010-11-22T11:50:52 some* 2010-11-22T11:51:23 *** narnach1 has joined #aichallenge 2010-11-22T11:51:23 *** narnach1 is now known as narnach_mb 2010-11-22T11:51:25 *** lavalamp_ has joined #aichallenge 2010-11-22T11:53:52 ok thx :) do you know how fast the google machines are? 2010-11-22T11:56:05 google machines? 2010-11-22T11:57:29 the pc on which fights will be run 2010-11-22T11:57:35 or server 2010-11-22T11:57:36 *** McLeopold has joined #aichallenge 2010-11-22T11:59:58 I don't know whether that info is available yet. The current servers that are running the contests handle something like 10 games per second each. 2010-11-22T12:00:17 (there is a page hidden somewhere on the website showing the stats, but I can't find it -_-) 2010-11-22T12:00:23 @gamestats 2010-11-22T12:00:24 McLeopold: gamestats = http://ai-contest.com/game_stats.php 2010-11-22T12:00:28 the servers running now are the ones that will be used 2010-11-22T12:00:48 google doens't really have a whole lot to do with the contest 2010-11-22T12:00:58 their names on it and they provided some money this time 2010-11-22T12:00:58 ah ok. That's massively faster than your average home pc I guess? 2010-11-22T12:01:05 Mine certainly can't handle 20 games per second 2010-11-22T12:01:13 oh per minute 2010-11-22T12:01:18 nvm reading the charts wrong -_- 2010-11-22T12:01:27 it depends which bots are playing 2010-11-22T12:01:30 It could probably handle 20 RageBot games per minutes. 2010-11-22T12:01:41 yeah per minute is ok 2010-11-22T12:01:44 was reading it wrong 2010-11-22T12:01:50 my pc does about 1 game per 2-3 seconds 2010-11-22T12:01:58 (crappy bot :p) 2010-11-22T12:02:06 i hope it is fast i need much operatin power :D 2010-11-22T12:02:07 I need to improve my codes speed, apparently. 2010-11-22T12:02:32 but i hope i dont have to use assembler ^^ 2010-11-22T12:04:35 what are you doing that requires so much time per turn? 2010-11-22T12:04:42 does the contest have a deeper sense 2010-11-22T12:04:55 i'm beginning to have doubts about my current method of making the game tree 2010-11-22T12:04:58 yes, this is Ender's Game except in real life 2010-11-22T12:05:00 predicting future states of all planets after every potential fleet order 2010-11-22T12:05:03 calculating turns in future 2010-11-22T12:05:04 it has its benefits and problems 2010-11-22T12:05:15 and after revoke orders 2010-11-22T12:05:16 like chess 2010-11-22T12:05:39 McLeopold: i think our code would be quite similar 2010-11-22T12:05:44 i expect most peoples will be 2010-11-22T12:05:58 I think my code is similar to almost everyones. Whoever has the best bugs wins. :) 2010-11-22T12:06:06 very true 2010-11-22T12:06:09 ^^ 2010-11-22T12:06:47 I had to run cProfile to find out why it was so slow. 2010-11-22T12:06:54 String parsing was 75% of it. 2010-11-22T12:07:10 When I take out the logging, I'll be alot faster. 2010-11-22T12:07:10 parsing the game state? 2010-11-22T12:07:38 i took out all the tokenize crap in the c++ starter package 2010-11-22T12:08:02 whois "testing" on tcp? 2010-11-22T12:08:03 i'm just inputting straight to a string stream and throwing out any comments 2010-11-22T12:08:08 cuz omg that game is taking loooooooooooong 2010-11-22T12:08:08 error323: me 2010-11-22T12:08:10 The second slowdown, in a 10 turn game, was about 39,000 calls to "how many ships do I have available" 2010-11-22T12:08:11 but it's broken 2010-11-22T12:08:19 i see 2010-11-22T12:08:21 So I added memoization 2010-11-22T12:08:34 I'm almost wishing I hadn't used the starter package now 2010-11-22T12:08:50 my code has almost nothing from the starter package 2010-11-22T12:09:04 I hung on to the starter Fleet and Planet objects for far too long 2010-11-22T12:09:16 Error323: did you like my terrible first move? :P 2010-11-22T12:09:25 did see it 2010-11-22T12:09:26 yet 2010-11-22T12:09:26 Yeah, you don't need fleet objects 2010-11-22T12:09:30 but fuck you won >_< 2010-11-22T12:09:41 :P 2010-11-22T12:09:45 link? 2010-11-22T12:10:02 http://72.44.46.68/canvas?game_id=1145125 2010-11-22T12:10:43 *** albert__ has joined #aichallenge 2010-11-22T12:10:45 the first move is an example of my doubts of my game tree 2010-11-22T12:11:00 what i think it's doing is making a move from me, having them attack it, then seeing their home planet as tackable 2010-11-22T12:11:06 what's up with the pathing from 2 to 10 to 6? 2010-11-22T12:11:45 hmmm but that's because its not good yet 2010-11-22T12:11:48 damnit 2010-11-22T12:11:58 antimatroid1: I guess it didn't see that he would land on his planets and be able to send back reinforcements? 2010-11-22T12:12:09 JamesMG: my redistribution minimuses distance to it's redistribution planet and it's closest enemy 2010-11-22T12:12:26 i don't even have the opponent expanding 2010-11-22T12:12:34 it doesn't work the way i'm doing my game tree 2010-11-22T12:12:53 i'm not really sure how to pick moves out otherwise though, it's too hard to do combinations 2010-11-22T12:13:43 although i haven't got it calculating closest enemies as i plan to yet either 2010-11-22T12:13:48 there's still quite a bit to be added 2010-11-22T12:13:54 and there's something broken with the tree 2010-11-22T12:14:33 bah, and it just crashed again 2010-11-22T12:15:34 McLeopold: I found out who testingFromPiter is 2010-11-22T12:15:42 it's accoun 2010-11-22T12:16:06 Yesssss: http://72.44.46.68/canvas?game_id=1145181 2010-11-22T12:16:45 *** McLeopold has quit IRC (Ping timeout: 245 seconds) 2010-11-22T12:16:52 http://72.44.46.68/canvas?game_id=1141932 i was proud of that before 2010-11-22T12:18:39 *** fawek has quit IRC (Remote host closed the connection) 2010-11-22T12:18:45 Top 10 players: bocsimacko(3384), Raschi(3281), bix0r4ever(3270), dmj111(3220), asavis(3166), _Astek_(3149), medrimonia(3139), luksian(3138), felixcoto(3125), smloh1(3117) 2010-11-22T12:22:21 *** bhasker has joined #aichallenge 2010-11-22T12:23:21 *** bhasker_ has joined #aichallenge 2010-11-22T12:24:59 mcleopold you planning on submitting a new version? 2010-11-22T12:25:32 I have got to fix this bug: http://72.44.46.68/canvas?game_id=1144863 2010-11-22T12:25:52 watch my route/drain/fortify code steal defenders away from that big neutral 2010-11-22T12:26:28 actually I did fix this bug. I fix so many bugs yesterday I broke it beyond all recognition and had to roll back. 2010-11-22T12:26:50 *** bhasker has quit IRC (Ping timeout: 265 seconds) 2010-11-22T12:26:55 now I'm merging my changes back in one at a time 2010-11-22T12:27:11 this bot is my current submission with a new eval function 2010-11-22T12:30:27 *** Accoun has quit IRC () 2010-11-22T12:32:51 so lavalamp a simple rule that works well 2010-11-22T12:32:52 for me 2010-11-22T12:33:02 is to not route ships from a planet thats under attack 2010-11-22T12:33:12 i meant not forward 2010-11-22T12:33:32 so ships pile up at that planet till the attack stops flowing 2010-11-22T12:34:16 *** McLeopold has joined #aichallenge 2010-11-22T12:35:34 McLeopold: you can take two planet safely on the first move for map 7 2010-11-22T12:35:43 the og map 7 2010-11-22T12:36:08 damnit got a reaaaaaaally nasty sigsegfault here 2010-11-22T12:36:15 can't find the source >_< 2010-11-22T12:36:27 my bot has something causing random crashes atm 2010-11-22T12:36:39 i've been trying to work out what's causing it 2010-11-22T12:39:10 bhasker_: In my case the problem is that my defense code didn't reserve the ships properly 2010-11-22T12:39:23 ah k 2010-11-22T12:41:46 *** Accoun has joined #aichallenge 2010-11-22T12:45:28 anyone see this on the forums? http://ai-contest.com/forum/viewtopic.php?f=19&t=1094&p=7552#p7552 2010-11-22T12:45:36 I can't check right now 2010-11-22T12:46:57 i'll check 2010-11-22T12:47:45 http://zeroviz.us:8080/canvas.cgi?game_id=20411 2010-11-22T12:47:49 looks right to me :\ 2010-11-22T12:51:34 *** Error323 has quit IRC (Read error: Operation timed out) 2010-11-22T12:52:41 *** Error323 has joined #aichallenge 2010-11-22T12:52:55 i love a good sigsegv 2010-11-22T12:54:13 *** McLeopold has quit IRC (Ping timeout: 240 seconds) 2010-11-22T12:54:41 i wasn't aware there was such a thing as a good sigsegv 2010-11-22T12:55:11 :P 2010-11-22T12:55:34 there is in the land of sarcasm 2010-11-22T12:56:26 *** McLeopold has joined #aichallenge 2010-11-22T12:56:31 anitmatroid1: sry, akf 2010-11-22T12:56:35 uh, afk 2010-11-22T12:56:45 no 2010-11-22T12:56:50 you can only take one 2010-11-22T12:57:55 no, you can take 2 2010-11-22T12:58:12 you can take the neutrals with 4 and 12 2010-11-22T12:58:16 map 7, with the 12+5 's in the center 2010-11-22T12:58:16 try with jbotmanager 2010-11-22T12:58:23 yes 2010-11-22T12:58:37 my bot takes the neutral with 12 and 4 2010-11-22T12:58:58 i thought it was a bug, but upon playing it in human mode, and sending everything at 12 from the first turn, it doesn't lose the planet 2010-11-22T12:59:49 oh, yeah, you can take those two no problem. I thought you meant both in the center. 2010-11-22T13:01:47 *** yasith_ has quit IRC (Ping timeout: 264 seconds) 2010-11-22T13:02:52 janzert: is currently server #0 hosting the website and computing the rankupdate? 2010-11-22T13:03:18 currently as in that is the way it has always been yes 2010-11-22T13:03:19 if so it might be a good idea to let another server compute the ranings 2010-11-22T13:03:21 rankigns* 2010-11-22T13:03:39 yeah, it'd be nice to offload that and I have thought about it 2010-11-22T13:03:47 but it's not completely trivial to do it 2010-11-22T13:04:33 cya all 2010-11-22T13:04:34 most likely not extremely difficult either, but still time I don't really have :) 2010-11-22T13:04:39 *** TEAM_AI has quit IRC () 2010-11-22T13:05:15 I see 2010-11-22T13:09:53 antimatroid1: I'm going to fix a bug, then show you a good map7 start. 2010-11-22T13:09:55 *** Mjothvitnir has joined #aichallenge 2010-11-22T13:10:31 Now that the current map set has been in play for around as long as the finals will run, how would everyone like a new set on the main site? 2010-11-22T13:11:13 McLeopold: my bot owns on map 7 :) 2010-11-22T13:11:17 as long as it is the same generator it works for me. 2010-11-22T13:11:32 put a link on pastewars 2010-11-22T13:11:35 @pastewars 2010-11-22T13:11:36 McLeopold: pastewars = http://visualizer.naktibalda.lt/ 2010-11-22T13:11:38 jaznert: put up like 1000 new ones 2010-11-22T13:11:44 so it's just like all new maps all the time 2010-11-22T13:11:51 lol janzert* 2010-11-22T13:12:01 hehe, nah I don't want to that far :) 2010-11-22T13:12:11 i don't know how to get the string 2010-11-22T13:12:13 janzert: put up crazy_fast maps :) 2010-11-22T13:12:18 can i from jbotmanager? 2010-11-22T13:12:35 I don't know if jbotm's save game is compatible 2010-11-22T13:12:40 McLeopold: you just want to get me mobbed right? 2010-11-22T13:12:43 I just run command line 2010-11-22T13:12:52 janzert: will it be any different? 2010-11-22T13:13:05 so much different 2010-11-22T13:13:12 people would jump up and down 2010-11-22T13:13:39 now they jump up and down, after crazy maps they would jump up and down on janzert 2010-11-22T13:14:07 don't do it janzert, we need you to hold everything together next contest too 2010-11-22T13:14:43 *** yasith_ has joined #aichallenge 2010-11-22T13:14:50 I probably won't be around for the next one, so might as well take the blame for this one right ;) 2010-11-22T13:15:31 haters gonna hate 2010-11-22T13:18:45 Top 10 players: bocsimacko(3379), bix0r4ever(3257), dmj111(3206), Raschi(3176), felixcoto(3162), asavis(3136), davidjliu(3122), _Astek_(3118), wagstaff(3114), medrimonia(3113) 2010-11-22T13:30:11 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716]) 2010-11-22T13:37:47 cool its working :D 2010-11-22T13:37:53 :) 2010-11-22T13:39:24 bah, i've had enough for tonight 2010-11-22T13:39:30 i'll keep working later 2010-11-22T13:47:11 You LOSE against pacifist 2010-11-22T13:47:14 damn 2010-11-22T13:47:19 thats humiliating 2010-11-22T13:53:26 bhasker: You autosniped yourself in the lower right: http://72.44.46.68/canvas?game_id=1146294 2010-11-22T13:53:46 oh wait nvm 2010-11-22T13:53:59 I didn't see that fleet under the other one 2010-11-22T13:54:53 still, sniping fail ^^ 2010-11-22T13:56:07 *** medrimonia2 has joined #aichallenge 2010-11-22T13:57:10 *** medrimonia has quit IRC (Ping timeout: 245 seconds) 2010-11-22T13:57:20 *** medrimonia1 has quit IRC (Ping timeout: 240 seconds) 2010-11-22T13:57:21 *** medrimonia has joined #aichallenge 2010-11-22T14:00:38 Accoun does not deal with poor opening moves: http://72.44.46.68/canvas?game_id=1146433 2010-11-22T14:01:22 A lot of my wins are because I did something suboptimal that the other guy didn't expect :/ 2010-11-22T14:05:12 if you win... was it a bad opening move ^o) 2010-11-22T14:06:33 I'm personally having a lot of trouble deciding what a "good" opening is. I almost always have worse expansion than my opponents 2010-11-22T14:06:43 *** Ice_Harley has joined #aichallenge 2010-11-22T14:07:01 I think that type of map is harder than others. 2010-11-22T14:07:12 Heuristic bots will do poorly on it. 2010-11-22T14:07:44 *** fawek has joined #aichallenge 2010-11-22T14:09:03 Do we still have bug, when player make invalid move it always goes to player #1? 2010-11-22T14:10:03 *** perror has quit IRC (Quit: Bye all !) 2010-11-22T14:10:41 http://ai-contest.com/visualizer.php?game_id=8326847 I make bot, who does moves same as Raftan and my bot doesn't make invalid move... 2010-11-22T14:12:07 *** narnach1 has joined #aichallenge 2010-11-22T14:12:07 *** narnach_mb has quit IRC (Read error: Connection reset by peer) 2010-11-22T14:12:54 *** perror has joined #aichallenge 2010-11-22T14:13:39 antimatroid: you still up 2010-11-22T14:13:45 antimatroid1 2010-11-22T14:17:54 map7 contest on tcp2.zeroviz.us, any takers? 2010-11-22T14:18:45 I would if I were at home.. 2010-11-22T14:18:47 Top 10 players: bocsimacko(3378), bix0r4ever(3323), Raschi(3280), dmj111(3212), _Astek_(3140), medrimonia(3124), felixcoto(3122), wagstaff(3118), davidjliu(3117), asavis(3114) 2010-11-22T14:19:25 It really seems that my bot is doing well on official 2010-11-22T14:20:34 I'm on McLeopold 2010-11-22T14:20:38 *** perror has quit IRC (Quit: Bye all !) 2010-11-22T14:20:38 k 2010-11-22T14:20:52 this bot probably won't do it right :) 2010-11-22T14:20:55 bam! 2010-11-22T14:21:19 who's next? 2010-11-22T14:21:42 I think I'm playing on it :p 2010-11-22T14:21:56 wow, that was a really bad opening move 2010-11-22T14:22:11 Mistmanov: port 996 2010-11-22T14:22:14 98.247.248.39 995? 2010-11-22T14:22:16 oh 2010-11-22T14:22:17 :p 2010-11-22T14:22:35 someone is playing on 995 btw.. 2010-11-22T14:22:45 yeah, it's been active for a while 2010-11-22T14:22:52 port 996 is crazy maps 2010-11-22T14:22:57 ewww 2010-11-22T14:23:09 okay, 2 for 2 2010-11-22T14:23:13 Mistmanov? 2010-11-22T14:23:16 that was my pseudo tree bot 2010-11-22T14:23:27 connected now I think 2010-11-22T14:24:09 hmm... I have more bugs 2010-11-22T14:24:23 but my opening was good enough 2010-11-22T14:25:10 my opening is really bad :( 2010-11-22T14:25:14 it alway is :( 2010-11-22T14:25:26 JamesMG: go again 2010-11-22T14:26:50 JamesMG? 2010-11-22T14:27:50 btw, do any of the crazy maps have more than 40 planets? if so, they would kill my bot 2010-11-22T14:27:57 41 :) 2010-11-22T14:28:04 seriously? -_- 2010-11-22T14:28:08 yes 2010-11-22T14:28:42 how come the game looks different? 2010-11-22T14:28:48 do you do random stuff? 2010-11-22T14:29:01 maybe i did? 2010-11-22T14:29:03 new maps in rotation 2010-11-22T14:29:09 I shouldn't 2010-11-22T14:29:11 link? 2010-11-22T14:29:23 so I can download to tcp 2010-11-22T14:29:37 just in svn right now 2010-11-22T14:29:42 one of us may have a planet id bias 2010-11-22T14:30:34 if 2 planets have the same "should I attack this" score, there will probably be a bias yes 2010-11-22T14:30:53 (sorting thing will then pick the one with the lowest ID ?) 2010-11-22T14:31:03 It's probably mine too 2010-11-22T14:31:26 that second game looked better for me 2010-11-22T14:31:37 I'm going to stop the map7 stuff 2010-11-22T14:31:42 So you can disconnect 2010-11-22T14:31:59 k 2010-11-22T14:34:05 hmmmm 2010-11-22T14:34:14 i've got timeouts :/ 2010-11-22T14:34:22 that's really )(*#$ 2010-11-22T14:35:04 New map set on zeroviz:995 2010-11-22T14:35:19 why? 2010-11-22T14:36:01 so you can find more bugs, of course 2010-11-22T14:36:35 New maps that are running on official 2010-11-22T14:36:41 Not ones I generated 2010-11-22T14:38:13 And available for download now... 2010-11-22T14:40:50 janzert, do we still have bug, when any player makes invalid move player #1 lose? 2010-11-22T14:41:56 Are you guys experiencing timeouts? 2010-11-22T14:42:00 or is this just local 2010-11-22T14:42:15 Ice_Harley: no 2010-11-22T14:42:31 Ice_Harley: it was only ever a visualizer reporting bug btw 2010-11-22T14:42:49 the game was correctly credited in the profile game list and for ranking 2010-11-22T14:42:49 not getting any timeouts here on the tcp server.. 2010-11-22T14:43:24 damn 2010-11-22T14:43:31 must be local then 2010-11-22T14:43:33 :/ 2010-11-22T14:43:33 can i see invalid move i did? 2010-11-22T14:45:01 it's not recorded, sorry 2010-11-22T14:45:14 well, thanks for reply 2010-11-22T14:45:23 Ice_Harley: http://spz.nofate.com/planetwars/visualizer/ allows you to turn official games into a map, that you can then replay locally 2010-11-22T14:45:36 (assuming your game engine supports fleets.. not all of them do) 2010-11-22T14:45:49 handy for exactly this situation =) 2010-11-22T14:46:11 i did reply that game locally. I even did bot, that make moves like my opponent, but i can't reproduce that invalid order ) 2010-11-22T14:46:47 ow ok 2010-11-22T14:47:03 I don't think the invalid order makes it to the game state recorded. 2010-11-22T14:47:29 MceLeopold: ? 2010-11-22T14:47:43 game state is "preturn" error is "inturn" 2010-11-22T14:48:01 i understand 2010-11-22T14:48:05 ah ok, but if you feed your bot the game state of preturn 2010-11-22T14:48:13 it should then produce the error right? 2010-11-22T14:48:23 right, but it doesn't 2010-11-22T14:48:28 then it must make the same decision to produce the move that casues the error 2010-11-22T14:49:12 move that causes the error would be in the next "game state" 2010-11-22T14:49:25 goddamnit, I keep getting timeouts >_< :( :( :( 2010-11-22T14:49:44 McLeopold: did you happen to run your statistic gathering script over the new set yet? 2010-11-22T14:49:57 ping to google is 10 ms on average :/ 2010-11-22T14:50:00 how is this possible 2010-11-22T14:50:10 no, at it doesn't really do much yet 2010-11-22T14:50:21 *** Appleman1234 has joined #aichallenge 2010-11-22T14:50:32 http://bitbucket.org/McLeopold/planetwars-mcmaps/src/d6584b138c4d/analyze_maps.py 2010-11-22T14:51:05 it just returns some averages 2010-11-22T14:51:10 ahh, thanks 2010-11-22T14:51:16 I'm focused on my bot right now. 2010-11-22T14:51:23 Mjothvitnir, http://ai-contest.com/visualizer.php?game_id=8326847. This is strange game even because I was 6th, when i played it and my opponent was 2200. Why we play together? :) And then that invalid order... 2010-11-22T14:51:49 yeah I'm just curious because the sampling of maps I looked at from the set looks like we may run toward the high end of planet number in this set 2010-11-22T14:51:51 ouch, that's got to hurt the elo 2010-11-22T14:52:10 my script will look at that, let me check 2010-11-22T14:52:21 but that's after only looking at a fairly small sample so far 2010-11-22T14:52:37 later... 2010-11-22T14:52:54 Ice_Harley: The pairings use a pareto distribution for ranking mismatch 2010-11-22T14:53:00 *** Azrathud has joined #aichallenge 2010-11-22T14:53:23 so there are occasional large distances between ranks 2010-11-22T14:53:31 thats ok 2010-11-22T14:53:56 McLeopold: yep, no problem I'll probably try and run it here in a bit 2010-11-22T15:00:21 I think tcp is overloaded ? 2010-11-22T15:00:29 i have no problem at leopold's server 2010-11-22T15:01:09 I have some draws on time on tcp 2010-11-22T15:01:21 pretty sure that's something to do with the server 2010-11-22T15:08:32 could everyone here join zeroviz.us ? :X 2010-11-22T15:08:59 I would if I were at home 2010-11-22T15:11:51 *** McLeopold has quit IRC (Ping timeout: 265 seconds) 2010-11-22T15:16:11 Error323: it's down? 2010-11-22T15:18:05 foucist: no it seems i'm the only one experiencing lots of timeouts 2010-11-22T15:18:07 :/ 2010-11-22T15:18:46 Top 10 players: bix0r4ever(3386), bocsimacko(3381), Raschi(3280), dmj111(3219), asavis(3124), medrimonia(3123), _Astek_(3115), wagstaff(3108), protocolocon(3102), felixcoto(3100) 2010-11-22T15:18:51 http://www.zerovis.us/ ? 2010-11-22T15:19:12 er 2010-11-22T15:19:19 s/s.us/v.us/ 2010-11-22T15:19:25 zeroviz.us is good 2010-11-22T15:19:36 its dhartmei's server that gives me timeouts 2010-11-22T15:19:46 ah 2010-11-22T15:19:58 LOTS of timeouts 2010-11-22T15:19:59 :/ 2010-11-22T15:20:06 http://72.44.46.68/getplayer?player=E323-19.1-7-g97176a6 2010-11-22T15:20:29 lol 2010-11-22T15:21:17 can you see when your bot starts/finishes calculating a turn? 2010-11-22T15:21:26 yes 2010-11-22T15:21:43 I take it that takes less than a second? 2010-11-22T15:22:06 in that case.. your internet sucks =) 2010-11-22T15:22:14 average of 0.0002s :( 2010-11-22T15:22:23 no its not my internet 2010-11-22T15:22:29 i don't have problems on zeroviz 2010-11-22T15:22:44 well as in, your internet doesn't like the tcp server 2010-11-22T15:22:51 ye 2010-11-22T15:22:55 not many other options I think =) 2010-11-22T15:23:02 well it has liked it all this time 2010-11-22T15:23:06 so it is kinda weird imo 2010-11-22T15:23:42 maybe the CIA is monitoring your connection or something <_< 2010-11-22T15:24:11 could use a proxy 2010-11-22T15:24:16 lavalamp_: you also got a timeout i see 2010-11-22T15:24:28 perhaps tcp is overloaded 2010-11-22T15:24:31 yeah :/ 2010-11-22T15:24:37 no idea why 2010-11-22T15:24:51 it's possible I genuinely timed out 2010-11-22T15:25:06 seems unlikely though 2010-11-22T15:25:10 @seen boegel 2010-11-22T15:25:10 Error323: boegel was last seen in #aichallenge 23 hours, 10 minutes, and 41 seconds ago: * boegel doesn't have time to look into that in depth 2010-11-22T15:25:27 well... I don't see another explanation 2010-11-22T15:25:33 tcp has quite some activity 2010-11-22T15:25:44 and the protocol isn't exactly optimal for net traffic 2010-11-22T15:28:10 *** tapwater has joined #aichallenge 2010-11-22T15:30:20 :/ 2010-11-22T15:31:58 well this is really fucked up 2010-11-22T15:32:06 cuz i think i build something awesome 2010-11-22T15:32:11 and now i can't test it l_l 2010-11-22T15:32:28 *** McLeopold has joined #aichallenge 2010-11-22T15:32:38 janzert: I've got some map stats 2010-11-22T15:32:55 for the first official set of 50: 2010-11-22T15:33:02 *** bhasker has joined #aichallenge 2010-11-22T15:33:18 count: 21.70 2010-11-22T15:33:19 growth: 67.36 2010-11-22T15:33:19 snipable: 0.42 2010-11-22T15:33:47 *** bhasker_ has quit IRC (Read error: Operation timed out) 2010-11-22T15:33:51 for the next set of 25: 2010-11-22T15:34:00 count: 21.76 2010-11-22T15:34:00 growth: 67.92 2010-11-22T15:34:00 snipable: 0.36 2010-11-22T15:34:10 Very consistent. 2010-11-22T15:34:18 McLeopold: just got back, you still want me to run on map7? 2010-11-22T15:34:24 No 2010-11-22T15:34:28 wow that's extremely consistent 2010-11-22T15:34:41 McLeopold: what is the set of 50 and 25? 2010-11-22T15:34:53 map_r538 and map_r538a 2010-11-22T15:34:54 just the new set of 75 broken into two groups? 2010-11-22T15:35:02 hmm... 2010-11-22T15:35:05 I must be confused 2010-11-22T15:35:08 I would not expect that level of consistency for samples that small 2010-11-22T15:35:23 http://ai-contest.com/forum/viewtopic.php?f=3&t=1094&start=80#p7559 2010-11-22T15:35:31 has the zip with all 75 new maps 2010-11-22T15:35:56 oh, I only did the last 25 of the new set 2010-11-22T15:35:57 hold on 2010-11-22T15:36:06 ahh, ok 2010-11-22T15:36:13 you and your stupid zero padding... 2010-11-22T15:36:23 heh, sorry 2010-11-22T15:36:46 it makes them sort nicer in lexicographic sorts though :) 2010-11-22T15:37:13 okay, second set: 2010-11-22T15:37:14 count: 22.15 2010-11-22T15:37:15 growth: 71.45 2010-11-22T15:37:15 snipable: 0.53 2010-11-22T15:37:41 snipable really should be "unsnipable", which means growth >= num_ships 2010-11-22T15:38:11 and I'm not doing stdev yet 2010-11-22T15:38:49 Hmm appears my home internet connection went down 2010-11-22T15:39:40 McLeopold: I was going to ask about snipable... so 0.53% of planets are unsnipable or? 2010-11-22T15:39:43 *** choas has joined #aichallenge 2010-11-22T15:39:55 yes 2010-11-22T15:40:15 or, there is an average of 1 planet every other map that is unsnipable 2010-11-22T15:40:16 McLeopold how is snipable defined? 2010-11-22T15:40:26 ^^ growth >= num_ships 2010-11-22T15:40:52 I only tossed that in there so I could find maps I wanted to test on. 2010-11-22T15:41:23 sense most planets come in pairs, it is more like every 4th map has 2 2010-11-22T15:41:30 any so actually we still ended up just under what should be the long term average for planets per map 2010-11-22T15:41:31 since 2010-11-22T15:42:02 22.5? 2010-11-22T15:42:18 there is the check for an odd number of planets 2010-11-22T15:42:33 *** yasith_ has quit IRC (Ping timeout: 240 seconds) 2010-11-22T15:42:37 the even numbers are at the top 2010-11-22T15:42:40 15-30 2010-11-22T15:42:50 so I would expect it to drop to 21.5 2010-11-22T15:43:09 hmm, yeah for radial maps anyway 2010-11-22T15:43:12 *** phreeza_ has joined #aichallenge 2010-11-22T15:43:27 or is it both types 2010-11-22T15:43:32 I guess it is 2010-11-22T15:43:33 just radial 2010-11-22T15:43:58 but line will still have it's fair share of odd planet maps 2010-11-22T15:43:59 3/4 should be odd, 1/4 should be even 2010-11-22T15:43:59 heh, only a week and I've already forgotten the details 2010-11-22T15:44:13 Mjothvitnir: yep 2010-11-22T15:44:36 maybe you should randomize the odd planet rounding? 2010-11-22T15:44:43 McLeopold: where can i find the new maps? 2010-11-22T15:44:45 sometimes go up, sometimes go down 2010-11-22T15:44:56 http://tcp.zeroviz.us/maps.zip 2010-11-22T15:45:02 @maps 2010-11-22T15:45:02 McLeopold: Run as fast as you can and don't look back. 2010-11-22T15:45:06 nah, I'm not touching it 2010-11-22T15:45:21 not going to be a big deal either way 2010-11-22T15:45:25 janzert: ok, but I'm only talking theory now 2010-11-22T15:45:46 yeah, if you cared about it, it would be easy to correct 2010-11-22T15:45:51 That's why I like my generator better. The center planet is optional. :) 2010-11-22T15:46:39 *** phreeza has quit IRC (Ping timeout: 265 seconds) 2010-11-22T15:47:22 janzert: any other map stat you care about? 2010-11-22T15:48:00 planet number median and possibly mode might be interesting as well 2010-11-22T15:48:38 and some actual density measurements, clustering and such but those get a lot harder to measure 2010-11-22T15:48:51 dang, if I was at home I could have spit those out already 2010-11-22T15:49:39 density always is at the center for a radial map 2010-11-22T15:49:47 speaking of which, it is time to head there 2010-11-22T15:49:52 *** Mjothvitnir has quit IRC (Quit: Page closed) 2010-11-22T15:49:56 maybe a histogram of the distance table? 2010-11-22T15:50:15 density as defined by number of planets / area of smallest containing convex hull 2010-11-22T15:50:23 *** Ice_Harley has quit IRC (Quit: Page closed) 2010-11-22T15:50:38 yeah, that will have to wait... :) 2010-11-22T15:50:47 McLeopold: yeah that would be interesting 2010-11-22T15:51:24 oh, I should get a max distance average 2010-11-22T15:52:26 yes, max distance averages and overall averages would certainly be nice to see and fairly easy to produce 2010-11-22T15:52:59 it's hard to judge overall distance average 2010-11-22T15:53:17 given (1,5,9) and (4,5,6), how do you compare? 2010-11-22T15:53:41 clearly one is clustered and the other not 2010-11-22T15:54:23 yeah, it's not as meaningful as the max distance average 2010-11-22T15:54:31 if it has any use at all I guess 2010-11-22T15:55:13 maybe if we only analyzed planets on one half of the map? 2010-11-22T15:55:34 *** yasith_ has joined #aichallenge 2010-11-22T15:55:46 then the twin planet wouldn't keep pulling us back to the middle 2010-11-22T15:57:46 the median may be more interesting for overall average than the mean 2010-11-22T15:58:06 median is the one in the center? 2010-11-22T15:58:31 yes, if I understand you right :) 2010-11-22T15:58:57 it would give a sort of clustering measure I think 2010-11-22T15:59:38 the histogram would be pretty clearly better than either 2010-11-22T16:00:31 avg count: 22.15 2010-11-22T16:00:31 mid count: 23.00 2010-11-22T16:00:31 avg growth: 71.45 2010-11-22T16:00:31 mid growth: 70.00 2010-11-22T16:00:31 avg snipable: 0.53 2010-11-22T16:00:31 mid snipable: 0.00 2010-11-22T16:00:36 ^^ for new 75 set 2010-11-22T16:01:54 hmm, so a few low planet number maps are pulling down the average 2010-11-22T16:02:16 *** yasith_ has quit IRC (Ping timeout: 252 seconds) 2010-11-22T16:02:44 mean that is 2010-11-22T16:07:36 *** Cold-Phoenix has quit IRC () 2010-11-22T16:07:37 janzert: http://pastebin.com/fSB0UZ1a 2010-11-22T16:08:30 you can see that the odd number planet maps have spikes. :) 2010-11-22T16:08:59 growth looks pretty good, evenly spread 2010-11-22T16:10:08 so it is the drop from an even number to an odd number that is pulling down the average 2010-11-22T16:10:19 yep, even number planets seems a little low? but overall looks pretty good 2010-11-22T16:10:31 which makes for slightly fast games? :) 2010-11-22T16:11:36 *** p4p4_ has joined #aichallenge 2010-11-22T16:11:54 McLeopold: you seem like a fan of sudden death AI contest 2010-11-22T16:13:11 ? 2010-11-22T16:13:25 *** p4p4 has quit IRC (Ping timeout: 245 seconds) 2010-11-22T16:13:27 with map 7 and crazy_fast maps and all 2010-11-22T16:13:58 http://pastebin.com/RhPwnjc7 2010-11-22T16:14:05 there is the original set 2010-11-22T16:14:28 McLeopold: is that mapset an integration with the official maps? 2010-11-22T16:14:38 on your server 2010-11-22T16:14:41 yes 2010-11-22T16:14:49 mmmm ok 2010-11-22T16:15:04 I run the official maps plus 50 generated myself. 2010-11-22T16:15:12 with the official generator 2010-11-22T16:15:20 ok 2010-11-22T16:15:26 *** yasith_ has joined #aichallenge 2010-11-22T16:16:31 janzert: the original maps had more of a bell curve with the total growth of the map 2010-11-22T16:17:04 I wonder if that is due to being set at 23 planets? 2010-11-22T16:18:47 Top 10 players: bix0r4ever(3406), bocsimacko(3385), Raschi(3324), dmj111(3233), medrimonia(3131), asavis(3130), _Astek_(3114), wagstaff(3109), davidjliu(3101), protocolocon(3100) 2010-11-22T16:18:50 McLeopold: the growth range is a lot larger with a bit of a smaller sample so it's not surprising 2010-11-22T16:19:22 in other words, yes :) 2010-11-22T16:22:12 *** delt0r___ has joined #aichallenge 2010-11-22T16:22:44 *** McLeopold has quit IRC (Quit: Leaving) 2010-11-22T16:23:20 *** delt0r has quit IRC (Ping timeout: 240 seconds) 2010-11-22T16:29:47 yeah my connection is just screwed up hard. 2010-11-22T16:31:33 *** Appleman1234 has quit IRC (Ping timeout: 240 seconds) 2010-11-22T16:31:40 *** Appleman1234_ has joined #aichallenge 2010-11-22T16:34:43 *** Rubicon-|-Cross has joined #aichallenge 2010-11-22T16:36:25 *** p4p4_ has quit IRC (Quit: ChatZilla 0.9.84 [SeaMonkey 2.0a3/20090223135443]) 2010-11-22T16:36:41 I wasn't expecting playing a great opening to be so hard... 2010-11-22T16:38:57 *** amriedle has joined #aichallenge 2010-11-22T16:42:22 *** Palmik has quit IRC (Remote host closed the connection) 2010-11-22T16:42:39 *** AlliedEnvy has quit IRC (Remote host closed the connection) 2010-11-22T16:43:31 *** McLeopold has joined #aichallenge 2010-11-22T16:44:32 *** AlliedEnvy has joined #aichallenge 2010-11-22T16:46:34 *** AlliedEnvy has quit IRC (Remote host closed the connection) 2010-11-22T16:49:03 *** AlliedEnvy has joined #aichallenge 2010-11-22T17:02:25 *** Naktibalda has joined #aichallenge 2010-11-22T17:12:11 *** albert__ has quit IRC (Quit: Page closed) 2010-11-22T17:12:11 *** lavalamp_ has quit IRC (Quit: Page closed) 2010-11-22T17:17:28 *** Mistmanov has quit IRC (Quit: Page closed) 2010-11-22T17:17:47 *** yellowseed has quit IRC (Ping timeout: 276 seconds) 2010-11-22T17:18:48 Top 10 players: bocsimacko(3394), bix0r4ever(3383), Raschi(3340), dmj111(3252), asavis(3166), davidjliu(3128), medrimonia(3126), protocolocon(3114), wagstaff(3110), luksian(3106) 2010-11-22T17:25:12 *** flag has joined #aichallenge 2010-11-22T17:26:09 *** AlliedEnvy has quit IRC (Remote host closed the connection) 2010-11-22T17:27:21 *** AlliedEnvy has joined #aichallenge 2010-11-22T17:35:24 *** choas has quit IRC (Quit: leaving) 2010-11-22T17:38:37 *** Cold-Phoenix has joined #aichallenge 2010-11-22T17:39:49 *** Epona666 has joined #aichallenge 2010-11-22T17:42:02 *** jaspervdj has quit IRC (Quit: NEVER GONNA GIVE YOU UP NEVER GONNA LET YOU DOWN) 2010-11-22T17:53:41 *** narnach1 has quit IRC (Quit: Leaving.) 2010-11-22T17:58:20 *** yasith_ has quit IRC (Ping timeout: 240 seconds) 2010-11-22T18:09:27 janzert, I was thinking about the map generator. 2010-11-22T18:09:54 with a guarenteed center planet, you'll never have a twin planet in the center 2010-11-22T18:10:23 Imagine 2 growth 3 planets 1.03 distance right in the middle 2010-11-22T18:10:34 antimatroid: ^^ 2010-11-22T18:10:51 it's a shame :( 2010-11-22T18:11:08 *** yasith_ has joined #aichallenge 2010-11-22T18:12:46 anyone with linux here that's willing to share a binary? 2010-11-22T18:12:55 s/share/trade 2010-11-22T18:13:03 *** nullkuhl has joined #aichallenge 2010-11-22T18:14:38 *** JamesMG has quit IRC (Quit: Leaving.) 2010-11-22T18:15:21 i will share you my 0.23 code:) 2010-11-22T18:15:27 i am done with the contest 2010-11-22T18:15:35 you can play with it if you want 2010-11-22T18:15:39 its a pile of mess ^ 2010-11-22T18:15:44 so you probably don't want to look at the code 2010-11-22T18:15:48 :) 2010-11-22T18:16:28 in all the games I've watched of my bot, the only ones I lose, are against bots that defend from attacks and/or re-enforce their front lines 2010-11-22T18:16:51 so, I gota add in atleast one of those, to advance more then a few ranks :\ 2010-11-22T18:17:01 *** yasith_ has quit IRC (Ping timeout: 252 seconds) 2010-11-22T18:17:07 bhasker: ? 2010-11-22T18:17:30 i will mail you my code 2010-11-22T18:17:36 if you want a bot to test against 2010-11-22T18:17:45 i don't think iam going to develop it further 2010-11-22T18:17:48 *** fawek has quit IRC (Read error: Connection reset by peer) 2010-11-22T18:17:54 well i do, but mmm giving out your code is unwise 2010-11-22T18:17:55 *** amstan_ has joined #aichallenge 2010-11-22T18:17:55 *** ChanServ sets mode: +o amstan_ 2010-11-22T18:18:02 i don't think you should do that. 2010-11-22T18:18:49 Top 10 players: bix0r4ever(3396), bocsimacko(3376), Raschi(3368), dmj111(3247), asavis(3169), davidjliu(3167), _Astek_(3142), medrimonia(3118), george(3111), protocolocon(3105) 2010-11-22T18:19:14 the top10 is getting tighter 2010-11-22T18:19:21 27th is the submission dead line and your bot is already better than mine 2010-11-22T18:19:23 do they play more similarly too? 2010-11-22T18:19:25 so if you do submit my bot:p 2010-11-22T18:19:27 you will lose 2010-11-22T18:19:33 *** amstan_ has quit IRC (Client Quit) 2010-11-22T18:19:39 *** amstan_ has joined #aichallenge 2010-11-22T18:19:39 *** ChanServ sets mode: +o amstan_ 2010-11-22T18:19:47 i could actually just send you the .pyc 2010-11-22T18:19:51 the compiled files 2010-11-22T18:20:03 bhasker: are they crossplatform or are you linux? 2010-11-22T18:20:03 since its just bytecode 2010-11-22T18:20:08 its python 2010-11-22T18:20:12 so .pyc is cross platform 2010-11-22T18:20:14 should work 2010-11-22T18:20:17 nice 2010-11-22T18:20:25 bhasker: do you want my binary? 2010-11-22T18:20:28 nah 2010-11-22T18:20:32 :/ 2010-11-22T18:20:36 i am done with this contest, too much of a time sink lol 2010-11-22T18:20:40 and i don't have more time for it 2010-11-22T18:21:17 I see 2010-11-22T18:21:33 mega1.respect++: http://ai-contest.com/forum/viewtopic.php?f=17&t=1142 2010-11-22T18:21:34 lol 2010-11-22T18:21:45 he has a fanclub now 2010-11-22T18:21:46 is it late saturday it's due: or late friday? 2010-11-22T18:24:40 yea his bot is freaking awesome 2010-11-22T18:24:43 #echo [topic remove 4] [topic remove 5] [topic remove 6] [topic add http://tinyurl.com/aichallengefinale] 2010-11-22T18:24:49 and depressing to watch as it destroys my bot 2010-11-22T18:24:51 @echo [topic remove 4] [topic remove 5] [topic remove 6] [topic add http://tinyurl.com/aichallengefinale] 2010-11-22T18:24:51 *** contestbot changes topic to "Official Google AI Challenge: http://ai-contest.com/ || Channel Logs: http://contestbot.hypertriangle.com/ || No stupid questions. Don't ask to ask. || Turn count WILL stay at 200! || Mailing List moving to the http://tinyurl.com/behindthescenesforum" 2010-11-22T18:25:02 k.. wtf? 2010-11-22T18:25:07 bosimacko is the proof that lisp is best for ai algorithms :P 2010-11-22T18:25:15 it makes me feel like the indian soccer team playing against argentina 2010-11-22T18:25:18 @topic remove 6 2010-11-22T18:25:18 amstan: Error: '6' is not a valid topic number. 2010-11-22T18:25:21 @topic remove 5 2010-11-22T18:25:21 *** contestbot changes topic to "Official Google AI Challenge: http://ai-contest.com/ || Channel Logs: http://contestbot.hypertriangle.com/ || No stupid questions. Don't ask to ask. || Turn count WILL stay at 200!" 2010-11-22T18:25:25 @topic remove 4 2010-11-22T18:25:25 *** contestbot changes topic to "Official Google AI Challenge: http://ai-contest.com/ || Channel Logs: http://contestbot.hypertriangle.com/ || No stupid questions. Don't ask to ask." 2010-11-22T18:25:33 @topic add http://tinyurl.com/aichallengefinale 2010-11-22T18:25:33 *** contestbot changes topic to "Official Google AI Challenge: http://ai-contest.com/ || Channel Logs: http://contestbot.hypertriangle.com/ || No stupid questions. Don't ask to ask. || http://tinyurl.com/aichallengefinale" 2010-11-22T18:29:08 *** yasith_ has joined #aichallenge 2010-11-22T18:31:13 bhasker: so, where is you pyc at? 2010-11-22T18:31:27 i can mail it to you 2010-11-22T18:31:30 haven't uploaded it anywhere 2010-11-22T18:31:40 mcleopold@gmail.com 2010-11-22T18:31:42 k 2010-11-22T18:32:06 sent 2010-11-22T18:32:10 let me know if it works 2010-11-22T18:32:18 can I run it on my tcp? 2010-11-22T18:32:18 MyBot.py the driver is a .py 2010-11-22T18:32:22 yea go ahead 2010-11-22T18:32:30 but it has no real logic 2010-11-22T18:32:34 the modules are all .pyc 2010-11-22T18:32:38 I though it was good? 2010-11-22T18:33:13 McLeopold: btw, this channel is logged and accessible through http, so you might not want to put emails in it 2010-11-22T18:33:23 oops 2010-11-22T18:33:35 no i mean the .py has no logic thats exposed 2010-11-22T18:33:39 gmail's got pretty good filters 2010-11-22T18:33:41 did everyone hear about china taking the lead in computing power? 2010-11-22T18:33:44 ya, sure 2010-11-22T18:33:46 lol 2010-11-22T18:34:09 Epona666: and by computing power you mean fastest linux comp? 2010-11-22T18:34:37 i think its called Tianhe-1 2010-11-22T18:34:41 yes 2010-11-22T18:35:49 bhasker: import error, bad magic number :( 2010-11-22T18:35:54 hmm weird 2010-11-22T18:35:56 it worked for e323 2010-11-22T18:36:07 run it as python ./MyBot.py 2010-11-22T18:36:16 which version of python are you using 2010-11-22T18:36:19 maybe thats the problem 2010-11-22T18:36:20 ? 2010-11-22T18:36:26 mine is running python 2.6 on mac 2010-11-22T18:36:34 so the pyc's are probably compiled for 2.6 2010-11-22T18:36:41 hmm, okay 2010-11-22T18:36:50 try it with 2.6 if you have it installed 2010-11-22T18:36:50 I"m on 2.7 2010-11-22T18:36:55 ah 2010-11-22T18:37:16 bad magic number probably looks like the .pyc carries the python version 2010-11-22T18:37:21 and its probably flunking because of that 2010-11-22T18:37:21 2.6.6? 2010-11-22T18:37:47 starevening-lm:minmax bhasker$ python --version 2010-11-22T18:37:47 Python 2.6.4 2010-11-22T18:37:47 s 2010-11-22T18:37:53 so 2.6.6 should work i think 2010-11-22T18:38:32 *** amstan_ has quit IRC (Remote host closed the connection) 2010-11-22T18:39:11 *** nullkuhl has quit IRC (Quit: Konversation terminated!) 2010-11-22T18:39:17 does anyone know who was on top of the charts last time? 2010-11-22T18:41:33 been bix and bocs mostly 2010-11-22T18:42:46 mcleopold did it work? 2010-11-22T18:43:01 I think. On my server it did. 2010-11-22T18:43:03 ah k 2010-11-22T18:43:20 But you don't handle the "go" test very well. :) 2010-11-22T18:43:26 whats the "go" test? 2010-11-22T18:43:38 run the bot, type go, wait for a reply of go 2010-11-22T18:43:45 ah don't send any state? 2010-11-22T18:43:51 it will probably flunk 2010-11-22T18:44:21 *** amriedle has quit IRC (Quit: Lost terminal) 2010-11-22T18:46:56 *** nullkuhl has joined #aichallenge 2010-11-22T18:47:53 anyone else want a copy of the bot drop me a line 2010-11-22T18:50:44 *** yasith_ has quit IRC (Ping timeout: 255 seconds) 2010-11-22T18:50:58 http://tcp.zeroviz.us/canvas.cgi?game_id=20570 2010-11-22T18:51:08 bhasker: which bot ? 2010-11-22T18:51:11 gotta fix the name 2010-11-22T18:51:34 geeze, and you got the EOF error too. :P 2010-11-22T18:51:38 thats fine :) 2010-11-22T18:51:43 haha rawline? 2010-11-22T18:51:47 yeah 2010-11-22T18:51:51 i never bothered fixing it 2010-11-22T18:51:53 <-lazy 2010-11-22T18:52:08 antbot is much stronger than 0.23 2010-11-22T18:52:13 i am not sure why you don't upload it 2010-11-22T18:52:15 to official 2010-11-22T18:52:53 it's almost ready 2010-11-22T19:03:17 *** yasith_ has joined #aichallenge 2010-11-22T19:04:13 *** wh1teside__ has quit IRC (Remote host closed the connection) 2010-11-22T19:08:31 so where do you think your bot will end up Mcleopold 2010-11-22T19:08:36 top 10/20/30? 2010-11-22T19:08:47 top 30 for sure 2010-11-22T19:08:51 I'm hoping top 10 2010-11-22T19:09:13 *** Epona666 has quit IRC (Ping timeout: 240 seconds) 2010-11-22T19:13:47 *** yasith_ has quit IRC (Ping timeout: 245 seconds) 2010-11-22T19:15:09 *** davidjliu has joined #aichallenge 2010-11-22T19:18:48 Top 10 players: bocsimacko(3381), Raschi(3376), bix0r4ever(3336), dmj111(3267), malex81(3183), asavis(3176), davidjliu(3176), _Astek_(3124), medrimonia(3119), george(3116) 2010-11-22T19:19:03 McLeopold: what's ur botname on tcp then? 2010-11-22T19:19:43 must...catch...dmj... 2010-11-22T19:21:10 aren't you like 4? 2010-11-22T19:22:53 McLeopold 2010-11-22T19:23:12 Error323: ^^ that was the answer to you questions 2010-11-22T19:23:17 ah ok 2010-11-22T19:23:19 antbot 2010-11-22T19:23:38 so how come you think top 30? 2010-11-22T19:23:41 I never got time to write OxBot. :( 2010-11-22T19:23:53 Well, I think I can beat you. 2010-11-22T19:23:56 is it doing that well? 2010-11-22T19:23:57 You're 38 2010-11-22T19:24:08 antbot is quite strong E323 2010-11-22T19:24:09 *** Utkarsh has quit IRC (Ping timeout: 265 seconds) 2010-11-22T19:24:14 play against it on mcleopolds' server 2010-11-22T19:24:20 if he runs it that is 2010-11-22T19:24:29 maybe later tonight 2010-11-22T19:24:46 McLeopold: what's ur rating on tcp then? 2010-11-22T19:25:11 Don't know 2010-11-22T19:25:32 so how do you know its strong then 2010-11-22T19:25:33 ? 2010-11-22T19:25:41 i wonder :) 2010-11-22T19:25:50 he runs it from time to time just not continuously 2010-11-22T19:25:50 By observing what it is capable of. 2010-11-22T19:26:05 and it kicks my bots butt quite regularly ^ 2010-11-22T19:26:11 Don't worry Error323, I think antbot is really a lot better than my bot 2010-11-22T19:26:20 ? really 2010-11-22T19:26:40 I think so, from what I remember and what I've seen 2010-11-22T19:26:46 i hate it when i start up a new version and it loses its first game 2010-11-22T19:26:54 yeah me 2 >_< 2010-11-22T19:27:00 and the second oen 2010-11-22T19:27:02 I'll be very surprised if I'm better ranked than you McLeopold 2010-11-22T19:27:17 and you're 9 2010-11-22T19:27:27 this space.invaders bot is quite interesting to watch 2010-11-22T19:27:30 its been improving 2010-11-22T19:27:33 now at 282 2010-11-22T19:27:39 *** yasith_ has joined #aichallenge 2010-11-22T19:27:39 Yes but I'll be loosing some ranks I think 2010-11-22T19:27:46 So, the problem is, my bot is very sensitive to map type. 2010-11-22T19:27:53 Error323: Beat you on tcp !!! 2010-11-22T19:28:08 medrimonia2: which version 2010-11-22T19:28:12 So, I can beat some of the best, but I lose to some not so good bots. 2010-11-22T19:28:16 19.1^^ 2010-11-22T19:28:23 ah :P 2010-11-22T19:28:40 nice going, 3 games all losses 2010-11-22T19:28:40 well im not sure which version is better yet :( 19.0 or 19.1 2010-11-22T19:28:44 i give up on my maximizing bot 2010-11-22T19:28:48 http://zeroviz.us:8080/canvas.cgi?game_id=13761 2010-11-22T19:29:07 I've modified the code of my first turn a lot, but I don't really know if that's very effective 2010-11-22T19:29:27 can't believe it loses to flavor 1@#!@#!@# 2010-11-22T19:30:38 * medrimonia2 is waiting for an opponent on zeroviz.us 2010-11-22T19:30:48 zeroviz is too slow:-\ 2010-11-22T19:30:51 too few connections 2010-11-22T19:31:13 I'll try benzedrine so 2010-11-22T19:31:33 I hope I won't have too much time outs 2010-11-22T19:32:49 i started up a version on zeroviz 2010-11-22T19:33:12 are you on tcp or tcp2? 2010-11-22T19:33:19 tcp: connected to 98.247.248.39:995, waiting for game 2010-11-22T19:33:19 You currently have -999999 Elo 2010-11-22T19:33:19 There are currently 1 connections and 0 games running 2010-11-22T19:33:20 ... 2010-11-22T19:33:22 i'm gonna get some sigarets, brb 2010-11-22T19:34:20 bhasker : I've just gone to benzedrine... =( 2010-11-22T19:34:26 ah 2010-11-22T19:34:32 i am on both 2010-11-22T19:34:37 same version on both 2010-11-22T19:34:42 not the 0.23 2010-11-22T19:34:52 this is a version that usually performs worse in my own tests:-\ 2010-11-22T19:35:00 i just want to see how it does against others 2010-11-22T19:35:04 Accoun: is my bot driving you insane already? ;) 2010-11-22T19:35:32 *** greghaynes has quit IRC (Ping timeout: 240 seconds) 2010-11-22T19:36:52 bhasker: you're losing to my old one :) 2010-11-22T19:36:56 i know lol 2010-11-22T19:37:03 won the first two 2010-11-22T19:37:05 lost the third 2010-11-22T19:37:23 raschi! a slightly somewhat legit number 1... 2010-11-22T19:37:38 medri do you want my 0.23 version to test locally 2010-11-22T19:37:42 @ranking 2010-11-22T19:37:43 McLeopold: Top 10 players: Raschi(3402), bocsimacko(3384), bix0r4ever(3350), dmj111(3260), malex81(3235), asavis(3187), davidjliu(3167), medrimonia(3126), _Astek_(3118), george(3113) 2010-11-22T19:37:53 *** Utkarsh has joined #aichallenge 2010-11-22T19:38:46 hmm 2010-11-22T19:38:52 nice finish on this game http://zeroviz.us:8080/canvas.cgi?game_id=20575 2010-11-22T19:38:53 raschi is becomming better ye 2010-11-22T19:39:12 bhasker: you can drop the 8080 2010-11-22T19:39:12 *** greghaynes has joined #aichallenge 2010-11-22T19:39:14 @tcp 2010-11-22T19:39:15 McLeopold: tcp = (#1) http://www.benzedrine.cx/planetwars/, or (#2) http://72.44.46.68/, or (#3) 98.247.248.39:995 (zeroviz.us), or (#4) http://tcp.zeroviz.us/, or (#5) http://tcp2.zeroviz.us/ (crazy!) 2010-11-22T19:39:18 ah k 2010-11-22T19:39:37 i'm gonna sync my attacks aswell I think, or try to 2010-11-22T19:39:43 and that'll be the last of it 2010-11-22T19:39:52 geeze, that's a big last step 2010-11-22T19:40:00 you should have done that a long time back 2010-11-22T19:40:05 *** amriedle has joined #aichallenge 2010-11-22T19:40:52 *** phreeza_ has quit IRC (Quit: Computer has gone to sleep) 2010-11-22T19:41:36 yeah 2010-11-22T19:41:39 :P 2010-11-22T19:41:50 Hmm my attacks sucks, I haven't change them for more than a month 2010-11-22T19:41:53 i didn't realize its power 2010-11-22T19:42:07 and its quite tricky 2010-11-22T19:42:17 yes its hard to get it right 2010-11-22T19:42:17 with the changing environment 2010-11-22T19:42:21 but when you do 2010-11-22T19:42:21 I'm more worried about expansion and all 2010-11-22T19:42:25 its a beautiful thing to watch 2010-11-22T19:42:29 indeed 2010-11-22T19:42:36 I'm more defensive 2010-11-22T19:42:41 But it should also be more effective 2010-11-22T19:42:49 yes it is 2010-11-22T19:42:56 cause enemy can't counter till all your ships launch 2010-11-22T19:43:01 makes it harder to snipe neutrals 2010-11-22T19:43:04 But I've still got a lot of issues I should fix, but there's not enough time remaining 2010-11-22T19:43:08 when you launch from multiple planets 2010-11-22T19:43:15 yes 2010-11-22T19:43:28 medri keep working :) you will get there 2010-11-22T19:44:25 The best with this contest is that... when you've finished one thing, you've two other things to do 2010-11-22T19:44:35 haha 2010-11-22T19:44:37 so true 2010-11-22T19:44:38 So you can't finish 2010-11-22T19:46:19 I don't even know if I'll be able to improve my bot until the deadline because... for each little modification there's a lot of hours of work to verify that everything's right 2010-11-22T19:47:05 bhasker: I don't think it makes it harder at all 2010-11-22T19:47:25 who cares how many planets the fleets come from? 2010-11-22T19:47:31 A snipe is a snipe 2010-11-22T19:48:17 and if you send 50 then 25.. at a 60.. the opponent is doing good to land 50 there inbetween teh 60 and 25 2010-11-22T19:48:58 well, better, but not that good 2010-11-22T19:49:17 it if you bring it down to the magical unsnipable range that's bad 2010-11-22T19:49:45 a lot of people are resubmitting these days^^ 2010-11-22T19:50:49 In fact I would like to play galcon turn by turn, assisted by my bot, there it would be interesting 2010-11-22T19:51:09 (I don't like galcon in real time) 2010-11-22T19:53:32 yeah 2010-11-22T19:53:52 It would be a really hard game 2010-11-22T19:53:52 I wish I could tell my bot what was an important planet, live, and it would do the details. 2010-11-22T19:54:21 Just click for the next target and have it stream, with defense turned on. 2010-11-22T19:54:21 Yes McLeopold: because sometimes... my bot is so stupid... 2010-11-22T19:55:00 I tried beating RageBot in human mode. It's very hard, too much going on. 2010-11-22T19:55:17 Even Ragebot is hard???? 2010-11-22T19:55:29 Once it gets more than one planet, yes 2010-11-22T19:55:41 Because you don't have a way to specify good fleet amounts 2010-11-22T19:55:53 Gosh I can't imagine playing against bocsimacko =P 2010-11-22T19:56:59 *** Utkarsh has quit IRC (Ping timeout: 264 seconds) 2010-11-22T19:58:11 I should have spent more time on the contest since the beginning 2010-11-22T20:01:59 i shouldn't have given up on minimax at the star 2010-11-22T20:02:02 start 2010-11-22T20:02:04 i had it 2010-11-22T20:02:06 I think that my two worst weaknesses are my attacks and the fact that when I've less growth than the opponent, I don't try anything special 2010-11-22T20:02:08 then thought, can't be done 2010-11-22T20:02:14 and then heard others were doing it 2010-11-22T20:02:19 -,- 2010-11-22T20:02:35 I don't even know what minimax is^ 2010-11-22T20:02:54 wiki it 2010-11-22T20:03:00 rlly simple in theory :) 2010-11-22T20:03:06 *** Utkarsh has joined #aichallenge 2010-11-22T20:06:37 *** rebelxt has joined #aichallenge 2010-11-22T20:08:07 medrimonia2: damn you, ur bot has beaten me on official 2010-11-22T20:08:08 >_< 2010-11-22T20:08:55 I don't understand why my bot is so strong on official... 2010-11-22T20:09:19 tcp is just a random fest imo 2010-11-22T20:09:42 yep but I'm loosing a lot more on tcp than on official I think 2010-11-22T20:09:56 Maybe I've just luck since I've updated... 2010-11-22T20:10:04 i don't know 2010-11-22T20:10:06 i have the same 2010-11-22T20:11:45 *** Utkarsh has quit IRC (Ping timeout: 245 seconds) 2010-11-22T20:11:58 i have a feeling ur start wins have a huge impact on tcp 2010-11-22T20:13:04 *** yasith_ has quit IRC (Ping timeout: 240 seconds) 2010-11-22T20:14:46 I don't know 2010-11-22T20:15:15 Error323: initial wins help a bit in short term, but elo normally stabilizes after 50-100 games it seems 2010-11-22T20:15:53 rebelxt: well my experiments tell otherwise 2010-11-22T20:16:30 Error323: what about long run? i.e. do initial wins still matter after 200+ continuous games? 2010-11-22T20:16:56 locally 19.0 is better than 19.0-1-gebc868e, but the latter is doing better with both of them running overnight 2010-11-22T20:17:21 however i also tested it under another name and then it didn't perform better 2010-11-22T20:17:23 *** Utkarsh has joined #aichallenge 2010-11-22T20:17:24 :S 2010-11-22T20:17:30 Error323: 19.0.1 might be better against distribution of opponents on TCP even though it's weaker in local testing 2010-11-22T20:17:50 i ran into this just recently too 2010-11-22T20:18:11 yeah, its scary 2010-11-22T20:18:49 Top 10 players: Raschi(3416), bocsimacko(3388), bix0r4ever(3351), dmj111(3265), malex81(3223), asavis(3190), davidjliu(3157), medrimonia(3137), _Astek_(3131), george(3111) 2010-11-22T20:18:50 i think 19.0.1 is better against strong bots, but weaker against weaker bots 2010-11-22T20:18:55 local testing can lead to bad results sometimes.. it did in my case 2010-11-22T20:19:05 so when it can't get out of local optima it gets stuck 2010-11-22T20:19:07 Error323: yeah, might be the case 2010-11-22T20:19:22 isnt that beautiful to watch http://72.44.46.68/canvas?game_id=1151414 2010-11-22T20:19:37 oh yeah :) 2010-11-22T20:20:16 *** Zoost has quit IRC (Ping timeout: 250 seconds) 2010-11-22T20:20:25 hmm this version isnt too bad 2010-11-22T20:20:31 let me start it up under a new name 2010-11-22T20:20:34 to see what elo it can get 2010-11-22T20:20:46 *** Zoost has joined #aichallenge 2010-11-22T20:21:26 mcleopold start up antbot on your server 2010-11-22T20:21:30 i want to see how my bot does against that 2010-11-22T20:21:36 its mostly beating your old bot now 2010-11-22T20:21:37 just a second 2010-11-22T20:22:24 looking for a QA engineer to test a new buggy bot... ;-) 2010-11-22T20:23:07 rebel you want my last submissions binary? 2010-11-22T20:23:09 to test against 2010-11-22T20:23:49 whoops, I'm on benzedrine 2010-11-22T20:24:04 bhasker: that could be very helpful since we are always very close. what OS do you compile on? 2010-11-22T20:24:09 McLeopold: that's a drug you know 2010-11-22T20:24:10 ;) 2010-11-22T20:24:19 python 2010-11-22T20:24:20 :) 2010-11-22T20:24:33 bhasker: heh i'm python too 2010-11-22T20:24:49 i can send you a zip with .pycs and a .py to run the bot 2010-11-22T20:24:54 i already sent to a few people 2010-11-22T20:24:59 give me a mail id 2010-11-22T20:24:59 *** yasith_ has joined #aichallenge 2010-11-22T20:26:05 there we go 2010-11-22T20:27:55 invalid order for you 2010-11-22T20:28:08 yea i need to track that bug down:( 2010-11-22T20:28:33 interesting i was almost winning 2010-11-22T20:28:35 bah 2010-11-22T20:30:44 ouch got sniped in the middle 2010-11-22T20:31:37 sniped again 2010-11-22T20:31:46 I'm gonna take a little sleep, good night all 2010-11-22T20:31:52 i really need to fix my neutral acquisition 2010-11-22T20:32:18 i have tried various heuristics none of the work well :-\ 2010-11-22T20:32:20 I let my bot playing on zeroviz 2010-11-22T20:33:08 ouch, take now 2010-11-22T20:33:20 you see McLeopold, I've lost against you^^ 2010-11-22T20:33:34 that's your official entry? 2010-11-22T20:33:58 not the one I've modified recently 2010-11-22T20:34:04 I'll put my official 2010-11-22T20:34:45 does anyone have a tar of the new maps 2010-11-22T20:35:00 a zip? 2010-11-22T20:35:04 zip/tar 2010-11-22T20:35:07 http://tcp.zeroviz.us/maps.zip 2010-11-22T20:35:09 sigh my bot is being dump 2010-11-22T20:35:10 :-\ 2010-11-22T20:35:16 going for 99 size neutrals in the middle 2010-11-22T20:36:02 medrimonia_submit on zeroviz McLeopold 2010-11-22T20:36:22 We'll see what will happen^^ 2010-11-22T20:36:36 uh oh 2010-11-22T20:36:57 that bug is on the to do list 2010-11-22T20:37:11 6 losses in a row and counting 2010-11-22T20:39:00 that's the second loss for me 2010-11-22T20:39:36 I'm not sustaining my attacks well. 2010-11-22T20:39:39 i shall test on new maps 2010-11-22T20:39:43 and see if i can make improvements 2010-11-22T20:39:48 I get the planet, then let go 2010-11-22T20:41:31 *** rebelxt has left #aichallenge 2010-11-22T20:41:40 *** McLeopold has quit IRC (Quit: Leaving) 2010-11-22T20:42:16 my bot gets sniped a lot on thes new maps 2010-11-22T20:43:11 sweet 2010-11-22T20:43:12 http://72.44.46.68/canvas?game_id=1151828 2010-11-22T20:43:19 synchronized neutral caps 2010-11-22T20:43:23 paying off aalready :D 2010-11-22T20:43:53 nice 2010-11-22T20:44:40 oh mmm 2010-11-22T20:45:15 i really need to stop losing to flavor 2010-11-22T20:45:16 lol 2010-11-22T20:46:46 lol 2010-11-22T20:47:32 its weird 2010-11-22T20:47:38 my bot can be some godo bots 2010-11-22T20:47:42 and then it goes and loses to flavor 2010-11-22T20:47:50 annoying as hell 2010-11-22T20:49:50 *** Utkarsh has quit IRC (Ping timeout: 255 seconds) 2010-11-22T20:51:58 *** Naktibalda has quit IRC (Remote host closed the connection) 2010-11-22T20:54:04 So it looks like boscimacko has this one :-O 2010-11-22T20:56:37 *** Utkarsh has joined #aichallenge 2010-11-22T20:59:43 man.. hope I get this stupid bug fixed soon :( 2010-11-22T20:59:52 * burny doesn't wana go back into the fireswamp 2010-11-22T21:00:29 *** bhasker has quit IRC (Quit: bhasker) 2010-11-22T21:01:01 bocs isn't even first right now? 2010-11-22T21:01:06 why would you think they're a sure win? 2010-11-22T21:01:59 He is in first, and he wins against the second place person something like 75% of the time 2010-11-22T21:02:16 Also, my comment was slightly made in jest 2010-11-22T21:02:32 lol, k 2010-11-22T21:02:59 *** Utkarsh has quit IRC (Read error: Connection reset by peer) 2010-11-22T21:04:17 woot, won a game.. looks like it's fixed :) 2010-11-22T21:07:49 flag: no jest needed 2010-11-22T21:07:55 it's got a very good chance of winning 2010-11-22T21:11:56 *** Utkarsh has joined #aichallenge 2010-11-22T21:17:49 whooot 2010-11-22T21:17:56 synchronization is awesome 2010-11-22T21:17:58 <3333333 2010-11-22T21:18:02 antimatroid1: yup 2010-11-22T21:18:19 But I didn't want anyone to lose hope ;) 2010-11-22T21:18:51 Top 10 players: bocsimacko(3391), bix0r4ever(3361), Raschi(3339), dmj111(3245), asavis(3212), medrimonia(3152), _Astek_(3136), davidjliu(3136), malex81(3135), luksian(3117) 2010-11-22T21:19:32 *** Utkarsh has quit IRC (Ping timeout: 255 seconds) 2010-11-22T21:19:32 there's been a couple new people on teh top10 each day lately 2010-11-22T21:19:51 http://72.44.46.68/canvas?game_id=1152286 turn 89 :D :D :D 2010-11-22T21:21:55 Error323: what am i looking at? 2010-11-22T21:22:17 planet 14 2010-11-22T21:23:16 oh, now i get you 2010-11-22T21:23:20 i've been doing that all along :P 2010-11-22T21:24:06 *** Utkarsh has joined #aichallenge 2010-11-22T21:24:50 yeah well i didn't ^^ 2010-11-22T21:28:18 i also need to do it for attacks though 2010-11-22T21:28:29 but thats a lot more difficult with my system 2010-11-22T21:28:38 *** Utkarsh has quit IRC (Ping timeout: 245 seconds) 2010-11-22T21:28:44 hmm, it's fairly easy for me 2010-11-22T21:28:59 my moves consist of vectors of fleets, then i just add fleets to them from my planets 2010-11-22T21:29:28 though also more rewarding probably 2010-11-22T21:33:30 *** Utkarsh has joined #aichallenge 2010-11-22T21:35:29 *** virdo has joined #aichallenge 2010-11-22T21:36:41 *** jmcarthur has quit IRC (Quit: WeeChat 0.3.2) 2010-11-22T21:37:03 *** Epona666 has joined #aichallenge 2010-11-22T21:37:52 *** Utkarsh has quit IRC (Ping timeout: 240 seconds) 2010-11-22T21:44:47 *** ColdPhoenix has joined #aichallenge 2010-11-22T21:48:20 *** Cold-Phoenix has quit IRC (Ping timeout: 276 seconds) 2010-11-22T21:50:04 *** bhasker has joined #aichallenge 2010-11-22T21:50:29 *** Cold-Phoenix has joined #aichallenge 2010-11-22T21:52:49 *** Utkarsh has joined #aichallenge 2010-11-22T21:54:47 *** ColdPhoenix has quit IRC (Ping timeout: 255 seconds) 2010-11-22T21:57:28 *** Cold-Phoenix has quit IRC (Ping timeout: 240 seconds) 2010-11-22T22:00:51 *** Cold-Phoenix has joined #aichallenge 2010-11-22T22:02:20 you gotta love this view: http://72.44.46.68/getplayer?player=E323-19.0-5-ge16d2ed 2010-11-22T22:07:14 *** Cold-Phoenix has quit IRC (Ping timeout: 255 seconds) 2010-11-22T22:09:07 *** Utkarsh has quit IRC (Ping timeout: 252 seconds) 2010-11-22T22:10:17 *** Cold-Phoenix has joined #aichallenge 2010-11-22T22:14:56 *** Utkarsh has joined #aichallenge 2010-11-22T22:18:50 Top 10 players: bocsimacko(3398), bix0r4ever(3365), Raschi(3345), dmj111(3252), asavis(3228), davidjliu(3156), medrimonia(3150), _Astek_(3137), malex81(3133), felixcoto(3130) 2010-11-22T22:19:15 *** Utkarsh has quit IRC (Ping timeout: 245 seconds) 2010-11-22T22:20:43 is the owner of "Crapbot" around? 2010-11-22T22:22:25 *** Rubicon-|-Cross has quit IRC (Quit: Leaving) 2010-11-22T22:22:45 aye 2010-11-22T22:22:51 Crapbot at your service 2010-11-22T22:23:31 *** Sylph has joined #aichallenge 2010-11-22T22:24:32 flag: its a thorn in my eye ;) 2010-11-22T22:24:35 good job 2010-11-22T22:24:48 Error323: thanks 2010-11-22T22:25:05 You were beating me for some time, I was formerly ObligatoryShampoo 2010-11-22T22:25:29 ahhhhhh 2010-11-22T22:25:37 you have massively improved 2010-11-22T22:25:42 :) 2010-11-22T22:25:45 *** Utkarsh has joined #aichallenge 2010-11-22T22:25:47 we were always wondering who that was 2010-11-22T22:25:54 flag: what's ur name @ official? 2010-11-22T22:26:03 FlagCapper 2010-11-22T22:27:48 fuck 2010-11-22T22:27:51 i hate the 200 time limit 2010-11-22T22:27:55 i would have beaten iouri then 2010-11-22T22:28:21 Who are you on tcp? 2010-11-22T22:28:58 http://72.44.46.68/canvas?game_id=1153188 2010-11-22T22:29:00 *** Rubicon-|-Cross has joined #aichallenge 2010-11-22T22:29:00 pbot atm 2010-11-22T22:29:03 it keeps sending invalid orders 2010-11-22T22:29:07 and i'm testing new things 2010-11-22T22:30:03 i had that game in the bad though :( 2010-11-22T22:30:19 Yeah you had twice his production 2010-11-22T22:30:23 At the end there 2010-11-22T22:30:47 i've no idea what's causing invalid orders :| 2010-11-22T22:31:10 worst comes to worst hopefully it'll be fixed at the end when i put checks in before making moves 2010-11-22T22:31:47 I just check if my move is valid before I make it, although I try to fix the problems before they happen anyway 2010-11-22T22:32:17 yeah i used to have those checks, but then my old bot was so buggy i left them out with my new one, that way i knew of and was forced to fix any errors 2010-11-22T22:32:50 i'm getting my hopes up for top 10 again 2010-11-22T22:32:59 this bot seems to be good when it isn't making really stupid moves 2010-11-22T22:33:08 if i can iron out the stupid moves it might have a chance 2010-11-22T22:33:32 Yeah, I'm hoping for a top10 as well 2010-11-22T22:34:32 Competition is pretty tight though 2010-11-22T22:34:46 yeah, top 2-30 is going to be very competitive 2010-11-22T22:34:52 *** Utkarsh has quit IRC (Ping timeout: 259 seconds) 2010-11-22T22:34:59 hahah, I like the 2-30 2010-11-22T22:35:04 1 has been decided ;) 2010-11-22T22:35:07 :P 2010-11-22T22:35:14 there is a subcompetition for second place 2010-11-22T22:35:50 The first place is there just as a benchmark against perfection 2010-11-22T22:38:16 i need to work out how to model the opponent expanding 2010-11-22T22:38:41 the way i currently make my tree, it's hard to have them do that and not try to then snipe moves they haven't actually made 2010-11-22T22:39:14 *** Utkarsh has joined #aichallenge 2010-11-22T22:39:17 Expanding is the hardest part of the game IMO 2010-11-22T22:39:43 yeah, that's alright, i mean having the opponent make expansionary moves in their nodes of the gametree 2010-11-22T22:40:09 Yeah I know what you mean 2010-11-22T22:40:53 I just try to see if I'd be able to defend an all-out attack my on my expansion 2010-11-22T22:41:25 each "move" in my game tree has a single destination 2010-11-22T22:41:31 then i just pick the best string of moves found 2010-11-22T22:42:27 So you try to calculate every possible move and pick the ones with the best chance of succeeding or what? 2010-11-22T22:42:49 atm i'm picking as many moves as possible per node 2010-11-22T22:43:19 in the middle of beating manwe :) 2010-11-22T22:43:29 it'll probably send a wrong move before the end though 2010-11-22T22:43:57 haha 2010-11-22T22:44:25 nup, got the muggle 2010-11-22T22:44:36 nice 2010-11-22T22:45:22 *** Utkarsh has quit IRC (Read error: No route to host) 2010-11-22T22:46:44 http://72.44.46.68/canvas?game_id=1153477 2010-11-22T22:46:48 he didn't even last 100 turns 2010-11-22T22:47:34 i like it when my bot plays more like that 2010-11-22T22:47:41 Yours plays really well actually 2010-11-22T22:47:54 bah, I lost like 400elo, by adding the ability to sent ships to planets en-route to a target 2010-11-22T22:48:14 the joys of minimax :) 2010-11-22T22:49:01 I should've probably re-written mine to use minimax a long time ago, I'm using the same basic crap I've been doing since the beginning of the contest 2010-11-22T22:49:16 hw basic? 2010-11-22T22:49:17 Cycle through all my planets, see what I can kill with them 2010-11-22T22:49:22 i'm going to try alpha beta in the next few days and see if it's better 2010-11-22T22:49:30 but i don't trust it for cropping trees as i make them 2010-11-22T22:49:35 lol flag, same here.. wht ELO ya at with that? 2010-11-22T22:49:43 On the main server? 2010-11-22T22:49:49 ya 2010-11-22T22:49:51 burny: he's top 100 pretty sure 2010-11-22T22:50:02 Well by some giant leap of chance I'm in 14th right now 2010-11-22T22:50:06 But I ussually hover around 40th 2010-11-22T22:50:20 and you don't defend against attacks? 2010-11-22T22:50:25 No, I do 2010-11-22T22:50:30 ok, lol 2010-11-22T22:51:01 I just think of it as killing future enemy planets :) 2010-11-22T22:51:22 at the start of my moves i do a sweep of planets and make all defensive moves i can 2010-11-22T22:51:23 *** Utkarsh has joined #aichallenge 2010-11-22T22:52:06 I'm actually starting to favor just trying to kill his planets when I can't defend something 2010-11-22T22:52:37 Sometimes it gives you a second chance at life if you get a favourable trade 2010-11-22T22:52:48 not often enough to do it always 2010-11-22T22:53:10 Maybe not, but if it's probably going to die anyway 2010-11-22T22:53:10 attacking is usualyl better than just losing though 2010-11-22T22:53:11 if you can pick when it's best to do, sure, but i wouldn't make it standard 2010-11-22T22:53:24 It also depends on the growth of the planet you're trading it for 2010-11-22T22:53:33 Before I was just always defending 2010-11-22T22:53:37 Now I give it an option 2010-11-22T22:53:53 yeah, my game tree should consider all options 2010-11-22T22:54:18 Yeah, that approach is much better IMO 2010-11-22T22:54:18 how do you make a program to make tactics? 2010-11-22T22:54:36 there are two problems with the way i do my game tree 2010-11-22T22:54:36 1. i don't consider them expanding, that can be bad on the first turn 2010-11-22T22:54:37 2. it tries to take their planets after i've made moves for them, even on the first turn 2010-11-22T22:55:12 Maybe program the idea of potential moves into it? 2010-11-22T22:55:16 i need to try and fix those two things, or make them less of a problem before the deadline 2010-11-22T22:55:22 So you only take them into account in certian situations 2010-11-22T22:55:28 *** lavalamp_ has joined #aichallenge 2010-11-22T22:55:35 yeah, it's hard to work out how to do that though 2010-11-22T22:55:51 lets see how it does against baron 2010-11-22T22:56:48 bah, i made so many stupid moves 2010-11-22T22:57:27 Are you watching the games before they end up on the server? 2010-11-22T22:57:36 yeah, i watch them as they play 2010-11-22T22:57:40 How? 2010-11-22T22:57:44 jBotManager 2010-11-22T22:57:49 http://ai-contest.com/forum/viewtopic.php?f=18&t=751 2010-11-22T22:58:32 *** Rubicon-|-Cross has quit IRC (Quit: Leaving) 2010-11-22T22:59:20 Thanks 2010-11-22T22:59:28 thank delt0r 2010-11-22T23:00:10 True 2010-11-22T23:00:13 *** Utkarsh has quit IRC (Ping timeout: 240 seconds) 2010-11-22T23:05:35 i like this game http://72.44.46.68/canvas?game_id=1153752 2010-11-22T23:06:30 *** Utkarsh has joined #aichallenge 2010-11-22T23:12:20 flag nice work with crapbot 2010-11-22T23:15:32 *** Utkarsh has quit IRC (Ping timeout: 240 seconds) 2010-11-22T23:18:50 thanks 2010-11-22T23:18:51 Top 10 players: bocsimacko(3403), bix0r4ever(3385), Raschi(3357), dmj111(3262), asavis(3225), medrimonia(3166), malex81(3142), davidjliu(3137), felixcoto(3123), george(3112) 2010-11-22T23:19:28 *** Epona666 has quit IRC (Ping timeout: 245 seconds) 2010-11-22T23:20:58 *** jmcarthur has joined #aichallenge 2010-11-22T23:26:26 *** narnach1 has joined #aichallenge 2010-11-22T23:26:26 *** narnach1 is now known as narnach_mb 2010-11-22T23:30:38 *** Utkarsh has joined #aichallenge 2010-11-22T23:30:43 *** Prillicy has joined #aichallenge 2010-11-22T23:30:52 hmm 2010-11-22T23:31:06 I told it to calculate in distance, but it's not :( 2010-11-22T23:31:56 I think my revamp where I rebuilt everything from scratch is actually worse than my first bot (well, not think, 8 out of 10 of the fights I'm testing are showing my original as winner) 2010-11-22T23:33:50 Prillicy: i started my new bot weeks ago 2010-11-22T23:33:53 took a while to get it better 2010-11-22T23:34:16 well, it took a few days just to get it to parse correctly 2010-11-22T23:35:35 *** Utkarsh has quit IRC (Ping timeout: 276 seconds) 2010-11-22T23:40:16 *** mceier has quit IRC (Quit: leaving) 2010-11-22T23:41:11 *** Utkarsh has joined #aichallenge 2010-11-22T23:43:56 *** narnach_mb has quit IRC (Quit: Leaving.) 2010-11-22T23:45:30 *** Utkarsh has quit IRC (Ping timeout: 245 seconds) 2010-11-22T23:50:16 *** Utkarsh has joined #aichallenge 2010-11-22T23:51:32 *** foucist has quit IRC (Quit: leaving) 2010-11-22T23:54:40 *** Utkarsh has quit IRC (Ping timeout: 240 seconds) 2010-11-22T23:56:16 Would anyone here know how you can run a bot from a (windows) batch file, and STDIN closing properly for the bot? Is there another solution for something like this? 2010-11-22T23:57:31 In short, when I run a bot from another folder, it uses current folder for bots execution... no problem for most part (log files, whatever)... I can run the bot from a batch file that changes the starting directory to the proper bot root folder, but... it seems the batch file gets the STDIN close/end, but the bot never does... and so it hangs... 2010-11-22T23:57:37 *** Epona666 has joined #aichallenge