2010-10-29T00:01:49 *** rebelxt_ has quit IRC (Quit: Page closed) 2010-10-29T00:08:13 *** hellman has quit IRC (Remote host closed the connection) 2010-10-29T00:09:48 *** TheMacca has joined #aichallenge 2010-10-29T00:12:28 Top 10 players: bocsimacko(3966), rsergio(3906), GreenTea(3905), rs5147(3817), dmj111(3798), felixcoto(3791), shangas(3752), protocolocon(3742), Hazard(3715), sin_sun(3710) 2010-10-29T00:41:27 *** TravisD has quit IRC (Read error: Connection reset by peer) 2010-10-29T00:43:51 *** TravisD has joined #aichallenge 2010-10-29T01:03:10 *** Accoun has quit IRC (Ping timeout: 276 seconds) 2010-10-29T01:10:05 *** sigh has quit IRC () 2010-10-29T01:10:30 *** seisatsu has joined #aichallenge 2010-10-29T01:10:47 *** Accoun has joined #aichallenge 2010-10-29T01:12:29 Top 10 players: rsergio(3934), bocsimacko(3925), GreenTea(3917), rs5147(3841), felixcoto(3809), dmj111(3808), shangas(3774), protocolocon(3756), Hazard(3738), Accoun(3730) 2010-10-29T01:15:40 *** Utkarsh has quit IRC (Ping timeout: 276 seconds) 2010-10-29T01:16:46 *** delt0r has joined #aichallenge 2010-10-29T01:17:25 *** delt0r___ has quit IRC (Ping timeout: 264 seconds) 2010-10-29T01:20:52 evening 2010-10-29T01:24:23 *** Utkarsh has joined #aichallenge 2010-10-29T01:29:57 *** virdo has quit IRC (Ping timeout: 240 seconds) 2010-10-29T01:31:25 *** Utkarsh has quit IRC (Ping timeout: 255 seconds) 2010-10-29T01:33:33 *** nibalize1 has quit IRC (Quit: leaving) 2010-10-29T01:33:45 *** virdo has joined #aichallenge 2010-10-29T01:39:04 *** Utkarsh has joined #aichallenge 2010-10-29T01:41:04 *** Accoun_ has joined #aichallenge 2010-10-29T01:41:40 *** Accoun has quit IRC (Ping timeout: 276 seconds) 2010-10-29T01:49:01 *** virdo has quit IRC (Ping timeout: 272 seconds) 2010-10-29T01:53:27 *** perestrelka has joined #aichallenge 2010-10-29T02:05:43 *** hornairs has quit IRC (Ping timeout: 276 seconds) 2010-10-29T02:12:30 Top 10 players: rsergio(3920), bocsimacko(3913), GreenTea(3905), rs5147(3831), felixcoto(3798), dmj111(3796), shangas(3765), protocolocon(3744), Hazard(3728), Accoun(3720) 2010-10-29T02:25:33 *** Palmik has joined #aichallenge 2010-10-29T03:06:51 *** Olathe has quit IRC (Ping timeout: 252 seconds) 2010-10-29T03:11:17 *** Gorovkhaia2 has joined #aichallenge 2010-10-29T03:12:29 *** aerique has joined #aichallenge 2010-10-29T03:12:31 Top 10 players: GreenTea(3911), rsergio(3911), bocsimacko(3855), rs5147(3806), dmj111(3788), felixcoto(3767), shangas(3737), protocolocon(3731), Hazard(3702), Accoun(3701) 2010-10-29T03:27:21 *** perror has joined #aichallenge 2010-10-29T03:36:22 *** narnach_mb has quit IRC (Quit: Leaving.) 2010-10-29T03:48:15 *** lavalamp_ has joined #aichallenge 2010-10-29T03:49:55 *** Florian95 has joined #aichallenge 2010-10-29T03:52:07 alright, i'm curious about how you guys are storing the ships at neutral planets, as the way i originally set doesn't actually work properly 2010-10-29T03:52:29 do you store the amount to take a ship and the "ships" at a planet at a point in time as two different things? 2010-10-29T03:53:13 i had it stored as one, and then added information about ships at other planets for extra amounts to send, but then if they have fleets arriving at a neutral without taking it inside the distance you could get there, the calculation is very very wrong 2010-10-29T03:56:40 anyone... 2010-10-29T04:01:26 oh god, i just realised anothed huge bug with move selection aha 2010-10-29T04:03:43 *** rangzen has joined #aichallenge 2010-10-29T04:11:01 *** boegel has joined #aichallenge 2010-10-29T04:12:31 Top 10 players: rsergio(3901), GreenTea(3901), bocsimacko(3844), rs5147(3808), dmj111(3776), felixcoto(3757), shangas(3728), protocolocon(3720), Accoun(3708), Hazard(3699) 2010-10-29T04:34:41 antimatroid: move_optimally() 2010-10-29T04:34:45 hehe 2010-10-29T04:36:45 *** Naktibalda has joined #aichallenge 2010-10-29T04:37:57 do you know what i mean? 2010-10-29T04:38:16 antimatroid: you need to store that separately 2010-10-29T04:38:32 yeah, bugger, more data to store/copy everywhere 2010-10-29T04:39:52 I keep arriving/departing/waiting (will arrive in the future)/held (part of some order) per planet per turn. Also (obviously) ship count and owner, but those are calculated from the others. 2010-10-29T04:41:01 i just had future planets with ships, unusedShips, and incoming/outgoing fleets 2010-10-29T04:41:23 *** Florian95 has quit IRC (Quit: Leaving...) 2010-10-29T04:42:07 *** Commandir has joined #aichallenge 2010-10-29T04:43:18 was something changed on the server: http://ai-contest.com/visualizer.php?game_id=6304020 ? 2010-10-29T04:44:22 ohhhh 2010-10-29T04:44:45 feel my pain :P 2010-10-29T04:45:56 its probably the slowness of server #0 2010-10-29T04:46:15 doh 2010-10-29T04:46:25 dunno, i just found a terrible bug in my code before that might cause the crashing 2010-10-29T04:46:48 it'd be nice if they narrowed that down for you... 2010-10-29T04:46:59 basically, when picking moves, my bot was potentially making negative moves 2010-10-29T04:47:36 but i don't think it would have executed them, so probably shouldn't cause crashing, hmmm 2010-10-29T04:48:33 i'm a bit surprised my bot even worked aha 2010-10-29T04:52:36 *** Utkarsh has quit IRC (Ping timeout: 240 seconds) 2010-10-29T04:58:29 *** Utkarsh has joined #aichallenge 2010-10-29T05:06:55 *** TheMacca has quit IRC (Quit: Page closed) 2010-10-29T05:12:32 Top 10 players: GreenTea(3916), rsergio(3912), bocsimacko(3855), rs5147(3817), dmj111(3789), felixcoto(3782), protocolocon(3742), shangas(3738), Accoun(3720), Hazard(3710) 2010-10-29T05:15:27 *** Florian95 has joined #aichallenge 2010-10-29T05:16:08 oh man, so many bugs i love it 2010-10-29T05:16:16 should mean a decent spike in ranking when they're removed 2010-10-29T05:16:50 *** Itkovian has quit IRC (Quit: Itkovian) 2010-10-29T05:17:05 *** Itkovian has joined #aichallenge 2010-10-29T05:18:54 *** Itkovian has joined #aichallenge 2010-10-29T05:19:22 antimatroid, what are you at right now? 2010-10-29T05:21:03 my move selection was completely retarded 2010-10-29T05:21:16 and my "amounts to claim planets" are often wrong 2010-10-29T05:21:28 then i still have my new macro idea in the works 2010-10-29T05:21:31 will see how it all goes 2010-10-29T05:21:40 oh ranking wise, it varies too much :P 2010-10-29T05:21:41 my logging rocks: http://pastebin.com/L7Tzcy8Y 2010-10-29T05:21:58 haha, that's awesome 2010-10-29T05:22:21 wow 2010-10-29T05:22:27 i might have to do that 2010-10-29T05:23:01 i'm so jealous right now aha 2010-10-29T05:23:24 * lavalamp_ evil laugh 2010-10-29T05:23:44 on the official site i'm right at the very bottom, but that's cause it is crashing on there 2010-10-29T05:23:47 i was top 30ish before that 2010-10-29T05:24:10 oh nice, are you on the tcp server? 2010-10-29T05:24:16 *** Bobng has joined #aichallenge 2010-10-29T05:24:16 not at this minute 2010-10-29T05:24:19 working on stuff 2010-10-29T05:24:26 i should really be studying for exams, but meh 2010-10-29T05:24:42 *** Bobng has quit IRC (Client Quit) 2010-10-29T05:24:48 yeah mine is too broken at the moment... 2010-10-29T05:25:07 *** Bobng has joined #aichallenge 2010-10-29T05:25:47 i have like 25 files too, but most of it i don't need to touch atm 2010-10-29T05:25:52 i'm working on the planetwars files 2010-10-29T05:58:27 *** perestrelka has quit IRC (Ping timeout: 245 seconds) 2010-10-29T06:03:39 *** perestrelka has joined #aichallenge 2010-10-29T06:12:33 Top 10 players: GreenTea(3917), rsergio(3911), bocsimacko(3860), rs5147(3819), dmj111(3790), felixcoto(3784), protocolocon(3743), shangas(3740), Accoun(3717), Hazard(3712) 2010-10-29T06:16:43 *** jaspervdj has joined #aichallenge 2010-10-29T06:17:30 antimatroid i have 15 files 2010-10-29T06:23:54 *** hellman has joined #aichallenge 2010-10-29T06:38:31 im assuming source files? 2010-10-29T06:38:38 yeah 2010-10-29T06:38:41 ^^ 2010-10-29T06:38:47 yeah 2010-10-29T06:39:23 some implement different strategies 2010-10-29T06:39:37 yep 2010-10-29T06:39:39 what language are you using? 2010-10-29T06:39:43 although my new idea will probably use most of them 2010-10-29T06:39:45 c++ 2010-10-29T06:39:48 you? 2010-10-29T06:39:52 ah ok, im using C 2010-10-29T06:39:57 so i guess you have per-class files 2010-10-29T06:40:11 i use structs and some files have more than one 2010-10-29T06:40:16 oh k 2010-10-29T06:40:18 but it's fairly split up 2010-10-29T06:40:23 ah 2010-10-29T06:40:41 i have 1990 lines 2010-10-29T06:40:55 my gamestate file has 1400 :P 2010-10-29T06:41:00 :o 2010-10-29T06:41:01 1613 here 2010-10-29T06:41:03 c++ 2010-10-29T06:41:42 sloccount 2010-10-29T06:42:34 amstan: its a bright day 2010-10-29T06:42:38 I want clouds!!!!!! 2010-10-29T06:42:40 =) 2010-10-29T06:43:04 it was quite warm here today 2010-10-29T06:43:08 we're heading in to summer :) 2010-10-29T06:43:26 well its cold and shitty here 2010-10-29T06:43:30 we are heading into winter 2010-10-29T06:43:31 sucked in 2010-10-29T06:43:41 i hate it when the opposite happens, the internet is so happy :P 2010-10-29T06:43:44 and i'm like FUUU guys 2010-10-29T06:43:45 but in the netherlands that doesn't matter, since summer == winter lol 2010-10-29T06:43:57 rain all day ^^ 2010-10-29T06:44:08 where are you at then? 2010-10-29T06:44:10 tasmania doesn't have the best weather 2010-10-29T06:44:18 1386 sloccount 2010-10-29T06:44:20 we're pretty much a large island afterall 2010-10-29T06:44:39 Total Estimated Cost to Develop = $ 38,062 2010-10-29T06:44:40 Error323: where do you live in NL? Delft here 2010-10-29T06:44:40 wow 2010-10-29T06:44:40 xD 2010-10-29T06:44:46 that really is the other side of the world 2010-10-29T06:44:55 aerique: hi, amsterdam 2010-10-29T06:45:31 Error323: too far for a quick beer :) 2010-10-29T06:45:37 hahaha 2010-10-29T06:45:47 my sister studies in delft too 2010-10-29T06:45:48 cool 2010-10-29T06:46:54 *** dhartmei has quit IRC (Read error: Connection reset by peer) 2010-10-29T06:49:05 well assuming u study there 2010-10-29T06:49:07 heh 2010-10-29T06:49:44 Error323: no, too old for that :) i work in the hague 2010-10-29T06:50:25 aerique: ah hehe :] 2010-10-29T06:56:17 its frustrating 2010-10-29T06:56:23 i created a bot that ranks around 100 2010-10-29T06:56:44 then updated and "improved" it worked better on tcp and against my old verison 2010-10-29T06:56:53 and now i don't get out of the 300 rankings 2010-10-29T06:57:05 hmmm 2010-10-29T06:57:13 like i created a strategy that only works good against higher ranks 2010-10-29T06:57:14 maybe it has some pathological cases against dumb bots? 2010-10-29T06:58:15 yea i think so 2010-10-29T06:58:37 could it be too conservative? 2010-10-29T06:59:12 mmm not sure 2010-10-29T06:59:54 i can't find anything sticking out when observing replays 2010-10-29T07:00:35 which frightens me a little :P 2010-10-29T07:07:18 ouch 2010-10-29T07:12:33 Top 10 players: GreenTea(3907), rsergio(3886), bocsimacko(3848), rs5147(3803), dmj111(3794), felixcoto(3769), shangas(3730), protocolocon(3725), Accoun(3704), Hazard(3685) 2010-10-29T07:20:28 *** other_donkey has joined #aichallenge 2010-10-29T07:20:52 *** mceier has joined #aichallenge 2010-10-29T07:21:24 *** other_donkey has quit IRC (Client Quit) 2010-10-29T07:26:15 *** irchs has joined #aichallenge 2010-10-29T07:34:34 *** irchs has quit IRC (Quit: irchs) 2010-10-29T07:39:49 i really wish we'd picked a different resolution strategy for neutral planets 2010-10-29T07:47:54 *** Florian95 has quit IRC (Ping timeout: 240 seconds) 2010-10-29T08:11:28 *** phreeza has joined #aichallenge 2010-10-29T08:11:54 *** hellman has quit IRC (Remote host closed the connection) 2010-10-29T08:12:34 Top 10 players: GreenTea(3906), rsergio(3877), bocsimacko(3839), rs5147(3791), dmj111(3788), felixcoto(3775), protocolocon(3741), shangas(3731), Accoun(3692), Hazard(3681) 2010-10-29T08:19:25 *** perestrelka has quit IRC (Ping timeout: 255 seconds) 2010-10-29T08:20:08 *** perestrelka has joined #aichallenge 2010-10-29T08:25:38 *** hellman has joined #aichallenge 2010-10-29T08:27:50 *** hellman has quit IRC (Remote host closed the connection) 2010-10-29T08:48:08 zzzzzzzzzz 2010-10-29T08:48:21 aaaaaaaaaa 2010-10-29T08:48:32 woke up to 5th place for an old bot, wtf? 2010-10-29T08:48:59 tcp or official? :) 2010-10-29T08:50:33 is there a way to get the full logs from the tcp server, not just the condensed version for the viewer? 2010-10-29T08:52:02 tcp 2010-10-29T08:52:19 Charmander.3 2010-10-29T08:54:01 phreeza somebody posted code to parse the condensed version on the forums 2010-10-29T08:54:35 eburnette_: cool 2010-10-29T08:54:48 i'll look for that 2010-10-29T08:58:34 aha http://codepad.org/V462WeFe 2010-10-29T08:58:40 thanks eburnette_ and dmj111 2010-10-29T09:03:42 the format looked absolutely cryptic to me... now i can make a bit of sense 2010-10-29T09:10:40 *** Florian has joined #aichallenge 2010-10-29T09:12:34 Top 10 players: GreenTea(3887), rsergio(3828), bocsimacko(3818), dmj111(3757), rs5147(3755), felixcoto(3739), protocolocon(3695), shangas(3694), Accoun(3656), Hazard(3655) 2010-10-29T09:15:33 *** delt0r___ has joined #aichallenge 2010-10-29T09:16:47 *** delt0r has quit IRC (Ping timeout: 245 seconds) 2010-10-29T09:19:17 *** seisatsu has quit IRC (Quit: Leaving) 2010-10-29T09:35:06 *** Azrathud has quit IRC (Ping timeout: 240 seconds) 2010-10-29T09:36:17 *** cfaftw has joined #aichallenge 2010-10-29T09:46:00 *** yasith has quit IRC (Ping timeout: 250 seconds) 2010-10-29T09:51:17 *** amastaneh has joined #aichallenge 2010-10-29T09:51:26 this is a test 2010-10-29T09:51:34 anybody's here from IRAN 2010-10-29T09:51:37 ? 2010-10-29T09:52:13 *** Azrathud has joined #aichallenge 2010-10-29T09:52:52 anybody's here from IRAN ? 2010-10-29T09:53:00 no 2010-10-29T09:53:18 *** Azrathud has quit IRC (Client Quit) 2010-10-29T09:58:16 *** yasith has joined #aichallenge 2010-10-29T10:04:19 Zannick: does the python tester send a EOF at the end? 2010-10-29T10:05:22 *** Appleman1234 has quit IRC (Remote host closed the connection) 2010-10-29T10:06:28 *** lavalamp__ has joined #aichallenge 2010-10-29T10:12:35 Top 10 players: GreenTea(3880), rsergio(3821), bocsimacko(3812), dmj111(3755), rs5147(3744), felixcoto(3743), protocolocon(3680), Accoun(3679), shangas(3676), Hazard(3642) 2010-10-29T10:14:09 oh dear, shoot me now 2010-10-29T10:21:47 *** Olathe has joined #aichallenge 2010-10-29T10:21:57 *** hellman has joined #aichallenge 2010-10-29T10:23:53 *bang* 2010-10-29T10:25:43 i have been checking my changes don't work with the wrong file all night :P 2010-10-29T10:25:48 worst mistake ever 2010-10-29T10:26:17 doh 2010-10-29T10:26:21 hate it when that happens 2010-10-29T10:26:35 i've never done it, worst 2010-10-29T10:28:41 *** JamesMG has joined #aichallenge 2010-10-29T10:30:18 lol xD 2010-10-29T10:30:23 i think i must have done that before 2010-10-29T10:41:34 *** demerzel_ has joined #aichallenge 2010-10-29T11:00:43 *** aerique has quit IRC (Quit: ...) 2010-10-29T11:00:46 *** rouli has joined #aichallenge 2010-10-29T11:01:28 *** Azrathud has joined #aichallenge 2010-10-29T11:05:39 oh god, things are so broken aha 2010-10-29T11:05:43 this is going to take a while 2010-10-29T11:05:43 *** bhasker has quit IRC (Quit: bhasker) 2010-10-29T11:07:42 i am almost to the point of having a bot again - i think this is my 3rd recode lol 2010-10-29T11:11:56 *** n9thbit has joined #aichallenge 2010-10-29T11:11:57 I really wish I'd found out about this when it started and not a month after :( 2010-10-29T11:12:37 Top 10 players: GreenTea(3857), bocsimacko(3815), rsergio(3789), dmj111(3734), felixcoto(3733), rs5147(3723), shangas(3660), Accoun(3658), protocolocon(3649), Hazard(3621) 2010-10-29T11:12:48 found out about what lavalamp__? 2010-10-29T11:12:58 the contest 2010-10-29T11:13:32 oh 2010-10-29T11:13:50 you feeling out of time? 2010-10-29T11:14:14 I'm not sure, it really depends on how long it takes to get my eval function working 2010-10-29T11:15:05 everything is working pretty well, it's just not optimizing for conditions that will let it win 2010-10-29T11:15:15 *** Azrathud has quit IRC (Quit: Meh) 2010-10-29T11:15:37 are you using payoff matrix? there's a lot of possibilities 2010-10-29T11:15:51 i have used zero game theory concepts just pure math x > y 2010-10-29T11:16:07 might need to step it up lol 2010-10-29T11:16:12 yes I have a payoff matrix 2010-10-29T11:16:30 based on what 2010-10-29T11:16:35 atm it plays a series of tournamentns between different goals to find the best one 2010-10-29T11:17:08 If I had more than a second it would do multiple ply but there's no way with this time limit 2010-10-29T11:17:44 it's totally possible i think 2010-10-29T11:18:06 my first bot ran every possible source/dest from 1 - numships before every move 2010-10-29T11:18:09 and it was in c# 2010-10-29T11:18:16 if i could do that in C# reliably i'm sure it's possible 2010-10-29T11:18:31 well, I will optimize after I get this working, I think I can get my node per second count back up. 2010-10-29T11:18:46 my move generator is not cheap at all, though, anymore 2010-10-29T11:18:58 i am trying to read all this game theory stuff but it's hard to see application in a bunch of formulas 2010-10-29T11:19:13 I just go off the diagrams 2010-10-29T11:19:15 ;) 2010-10-29T11:19:18 i am tempted to just stick with 'auto defenses' and reduce supply that way 2010-10-29T11:19:26 that's what i've been doing thus far 2010-10-29T11:19:34 lavalamp__ have you considered a kind of sparse tree sampling to get more ply? 2010-10-29T11:19:43 thats what i am trying to do 2010-10-29T11:19:47 but its not easy 2010-10-29T11:20:02 meaning each ply jumps ahead more than a turn? or something else? 2010-10-29T11:20:52 either that or not looking at every possibility within the ply 2010-10-29T11:21:17 just generate 1000 or so and then consider that ply full and move to the next 2010-10-29T11:21:26 what is a 'ply' to you? one turn launches? 2010-10-29T11:21:36 one ply for me is a move from each side 2010-10-29T11:21:37 and are you only considering one source/dest? 2010-10-29T11:21:44 ah 2010-10-29T11:21:46 my basic idea is to do tree width = log(ply) 2010-10-29T11:22:11 i am not dealing with formulas honestly - imo all moves can be reduced to 2 values 2010-10-29T11:22:20 rapidity of payback and overall value 2010-10-29T11:22:27 phreeza, my move generator makes a few goals and I calculate a few different implementations of them 2010-10-29T11:22:36 999999/ log(ships) blahblahblah just complicates it 2010-10-29T11:22:38 ah ok 2010-10-29T11:23:15 *** boegel has quit IRC (Quit: *poof!*) 2010-10-29T11:23:17 demerzel_, what do you state value in terms of? 2010-10-29T11:23:50 well i'm still playing with it - but i had a short estimator on whether this was going to 200 or not 2010-10-29T11:24:05 debugging my eval function has led me to do some cool logging: http://pastebin.com/L7Tzcy8Y 2010-10-29T11:24:06 if not i used a forecast of 20 + turns remaining/10 2010-10-29T11:25:00 my current bot is bugged in that it pauses for incoming but it's purely deterministic in value calculations 2010-10-29T11:25:14 how high did your last bot get? that's similar to what I was doing but I think it has limitations 2010-10-29T11:25:28 oh it's at 176 now with bug 2010-10-29T11:25:34 and 9-12 random C# crashes 2010-10-29T11:25:43 so not too shabby 2010-10-29T11:25:56 *** Cold-Phoenix has joined #aichallenge 2010-10-29T11:26:13 what's it do on the TCP server? go still doesn't run on the main server, grr 2010-10-29T11:26:15 it does have limitations 2010-10-29T11:26:35 beats the crap out of bogusbrainpower/flavor but loses to bosi/malazan 2010-10-29T11:26:45 then again everyone loses to mega1 haha 2010-10-29T11:26:59 the limitations are that it can't possibly calculate defense that way 2010-10-29T11:27:06 i just auto-calc what is necessary and then reduce supply 2010-10-29T11:27:09 ok mine does about the same 2010-10-29T11:27:42 yeah I need to add a goal of reducing the enemy's supply 2010-10-29T11:27:52 after seeing mega implement what i wanted (forecasted launches) i am doing the same in c++ now 2010-10-29T11:27:59 when I'm ahead 2010-10-29T11:28:06 nice thing about c++ is i don't have to think about optimizing ahead of time like i did in c# 2010-10-29T11:28:07 forecasted launches? 2010-10-29T11:28:16 yea watch his games he's got an order buffer 2010-10-29T11:28:24 like he'll launch 6 ships at p1 on turn 2 2010-10-29T11:28:35 follow it up with 12 more from a neutral he conquered on turn 6 2010-10-29T11:28:46 oh, yeah, I do that too 2010-10-29T11:28:58 i had an order buffer finished but never used it 2010-10-29T11:29:15 had to think about how to override it if necessary 2010-10-29T11:29:16 *** fawek has joined #aichallenge 2010-10-29T11:29:16 problem is I change my mind before issuing the second order 2010-10-29T11:29:21 i guess if your math is perfect you won't need to override it 2010-10-29T11:30:16 his choices are interesting i'd love to know what he's looking at 2010-10-29T11:30:27 like they're not always pure payback even with other planets forecasted 2010-10-29T11:30:45 accoun has a faster overall payback system for example 2010-10-29T11:31:29 yeah I think there are more important things than pure payback 2010-10-29T11:31:42 initially i would disagree 2010-10-29T11:32:01 but i'm not sure how to know when to make the transition 2010-10-29T11:32:25 hehe well if my stuff works out I will tell people what I did 2010-10-29T11:35:35 right now my bot assumes more is better and it causes it to take neutrals it shouldn't... 2010-10-29T11:36:05 *** amastaneh has quit IRC (Ping timeout: 265 seconds) 2010-10-29T11:36:32 raedwulf: no, it just terminates the bots 2010-10-29T11:40:19 has anyone heard from janzert recently? 2010-10-29T11:44:38 not for a couple days 2010-10-29T11:44:46 maybe he shot himself over trying to fix C# 2010-10-29T11:45:46 I might shoot myself if I do all this work and go never gets installed on the server :( 2010-10-29T11:47:06 that's why i switched to c++ 2010-10-29T11:47:12 *** fawek has quit IRC (Remote host closed the connection) 2010-10-29T11:47:25 plus i actually like c strings haha 2010-10-29T11:47:46 meh, I have no desire to write C++ in my spare time 2010-10-29T11:48:16 i can't use it anywhere 2010-10-29T11:48:37 this is nice tho: if(sscanf(inLine,"P %lf %lf %d %d %d",&x, &y, &owner, &ships, &growth) == 5) 2010-10-29T11:48:44 no tokenize streams blahblah shit 2010-10-29T11:49:26 ...except sscanf is a C thing... the C++ way to do that does indeed involve streams, hehe 2010-10-29T11:49:44 a holdover from C, I should say 2010-10-29T11:49:56 yea i write c++ like c with classes 2010-10-29T11:50:09 which is what it is ;) 2010-10-29T11:50:15 or started as at least 2010-10-29T11:50:19 so do my coworkers... :/ 2010-10-29T11:50:37 i hate every time i look up sscanf there's 20 posts about WTF DUDE USE STREAMS THATS WHAT C++ IS MAN 2010-10-29T11:51:10 or every time i look up arrays etc WTF DUDE USE VECTORS THATS WHAT C++ IS MAN 2010-10-29T11:51:17 I also find the ANSI file manipulation stuff much more convienient 2010-10-29T11:51:53 ...but vectors are better than arrays for many, many things 2010-10-29T11:52:14 http://72.44.46.68/canvas?game_id=766613 2010-10-29T11:52:15 wtf 2010-10-29T11:52:31 vectors are nice but if i wanted vectors i'd use them 2010-10-29T11:52:40 ever seen ship graph like that? 2010-10-29T11:52:56 it's like List in C# - i don't use it unless i really don't know the length, which is rare 2010-10-29T11:53:10 whoa that is an awesome graph 2010-10-29T11:53:35 it's almost perfect 2010-10-29T11:53:54 oh yeah, new heuristic bot is up 2010-10-29T11:54:03 interesting, h is implementing my ship draining code that I haven't written yet 2010-10-29T11:54:07 i'll start feeding it into the better stuff once i've cleaned it a bit more 2010-10-29T11:54:13 holy crap that's awesome 2010-10-29T11:54:55 do you like my finish :P 2010-10-29T11:55:11 yeah, one of those games than another 50 turns would've made more entertaining 2010-10-29T11:55:19 i'm not sure how it knew it was safe to do that without the tree's or game matrices 2010-10-29T11:55:32 not really, i was about to go in and take him out 2010-10-29T11:55:38 somebody was doing that - was it flavor or monty? where basically if they were ahead they'd just neutralize growth on all planets 2010-10-29T11:55:44 i tried it out but it's hard to get that far ahead 2010-10-29T11:55:50 without just destroying them anyway 2010-10-29T11:56:14 i have a different method, any unused ships each turn i redistrbute 2010-10-29T11:56:24 if i have no planets of mine to redistribute to, spam the closest enemy 2010-10-29T11:59:07 to what end though? there's no advantage in doing that 2010-10-29T11:59:17 i have no use for the ships 2010-10-29T11:59:21 if you both have 200 or 1200 it's actually more advantageous to hold the 1200 2010-10-29T11:59:28 bear in mind i am making future moves at each turn 2010-10-29T11:59:35 and reserving ships necessary to make them 2010-10-29T11:59:40 because if they launch their 1200 in a stupid or at least exploitable way having more is better 2010-10-29T12:00:15 it's doing much better than i expected given it's entirely heuristic based right now 2010-10-29T12:00:26 http://72.44.46.68/getplayer?player=h 2010-10-29T12:00:46 *** tapwater_ has joined #aichallenge 2010-10-29T12:02:30 http://72.44.46.68/canvas?game_id=766670 it failed for me there 2010-10-29T12:02:30 hmm my old C# bot is doing pretty good up to 150 now if i hadn't had like a dozen "crashes" it might have made it even with the bug 2010-10-29T12:03:37 *** tapwater has quit IRC (Ping timeout: 264 seconds) 2010-10-29T12:03:37 *** tapwater_ is now known as tapwater 2010-10-29T12:06:25 alright, i took spam the enemy out, see how that does 2010-10-29T12:07:27 your neutral choice needs work 2010-10-29T12:07:33 antimatroid, you can spam the enemy... just make sure your total production rate is higher 2010-10-29T12:08:03 why didn't it take the 29 2010-10-29T12:10:03 ignore my neutral choice atm 2010-10-29T12:10:14 it has a terrible hack in there cause it's using heuristics only 2010-10-29T12:10:32 it should do better once i throw it into different types of games and consider what the enemy can do as well 2010-10-29T12:11:57 *** dmj111 has quit IRC (Read error: Connection reset by peer) 2010-10-29T12:12:09 i am avoiding considering the enemy atm 2010-10-29T12:12:16 *** dmj111 has joined #aichallenge 2010-10-29T12:12:27 or rather the enemy's possible future moves 2010-10-29T12:12:36 Top 10 players: GreenTea(3859), bocsimacko(3822), rsergio(3790), dmj111(3738), rs5147(3728), felixcoto(3723), shangas(3667), Accoun(3665), protocolocon(3653), Hazard(3617) 2010-10-29T12:17:04 *** cfaftw has quit IRC (Quit: Leaving) 2010-10-29T12:19:14 ugh, just did a really complex piece of bit shifting and stuff... can't wait to debug it lol xD 2010-10-29T12:19:56 nice 2010-10-29T12:20:00 what about this requires bit shifting? 2010-10-29T12:20:06 i am a little worried about planets not actually counting 23 2010-10-29T12:20:09 i sorta rely on that 2010-10-29T12:20:17 demerzel_, me too 2010-10-29T12:20:28 i totally rely on it actually 2010-10-29T12:20:34 well, I'd be ok if there were fewer than 23 2010-10-29T12:20:35 my update function increments to 23 and back to 0 2010-10-29T12:20:44 I think 2010-10-29T12:21:30 i couldn't handle more than 40 or so i don't think 2010-10-29T12:21:32 not in one second 2010-10-29T12:21:58 well, i finding some neighbouring states, and im generating an index using bit shifting 2010-10-29T12:22:04 im* 2010-10-29T12:22:41 *** rouli has quit IRC (Quit: Page closed) 2010-10-29T12:22:45 at this point in the contest I think we have the right to be really upset if they change that on us... 2010-10-29T12:22:48 i have found the more complex i try to make my eval the more problems it has 2010-10-29T12:26:17 ea my bit function relies on 256 planets 2010-10-29T12:26:23 max 2010-10-29T12:26:25 :P 2010-10-29T12:26:26 I can't imagine why more complexity would add more bugs 2010-10-29T12:26:58 if the code has fixed sized structures.... 2010-10-29T12:26:59 I have found that the more time I spend on my bots the worse they do 2010-10-29T12:27:06 lol 2010-10-29T12:28:14 also, if your evaluation algorithm has something O(exp(n)) complexity with n = number of planets 2010-10-29T12:28:20 i can see why more planets=more headache 2010-10-29T12:28:22 lol 2010-10-29T12:28:55 the number of planets is directly proportional to the amount of headache 2010-10-29T12:28:59 I think my engine as a whole is probably at least O(n^4) 2010-10-29T12:29:13 i can't see a useful eval with log/cos/sin anything honestly it's just a disconnected number at that point 2010-10-29T12:29:54 I don't think I'm doing any O(exp(n)) anymore, I did used to have an O(exp(n)) for n = fleet count though... 2010-10-29T12:29:56 that was exciting. 2010-10-29T12:30:39 oh, my goal optimization algorithm is O(n!) 2010-10-29T12:31:00 ...I don't let that one run to completion... 2010-10-29T12:31:13 actually right now it's off completely 2010-10-29T12:31:18 i am still trying to make the leap from no goal beyond more points 2010-10-29T12:34:13 *** irchs has joined #aichallenge 2010-10-29T12:37:01 that doesn't even make sense 2010-10-29T12:37:18 i remove the declaration of a 3d vector that is not used from game state and my bot crashes 2010-10-29T12:39:27 lol, i don't even know how to debug that 2010-10-29T12:39:42 antimatroid: evilness 2010-10-29T12:40:11 antimatroid: there's a reason for that, you're overflowing your stack 2010-10-29T12:40:19 most probably 2010-10-29T12:40:49 how would changing the definiton of something cause that? 2010-10-29T12:41:42 variables take space up in the stack, if some other variable beside it overflows, it could induce crashes - but the crashes were prevented because they were stuffing things into that vector 2010-10-29T12:42:11 usually the compiler removes such definitions as they don't appear to be used, however 2010-10-29T12:42:16 is this vector a stl vector? 2010-10-29T12:42:28 or just a struct 2010-10-29T12:42:48 as in a mathematical vector 2010-10-29T12:42:48 it's a 3d vector of ints inside a struct 2010-10-29T12:42:53 ok 2010-10-29T12:42:58 stl vectors 2010-10-29T12:42:58 ew 2010-10-29T12:43:00 why 3d 2010-10-29T12:43:07 i have many of them 2010-10-29T12:43:14 do you have any nearby fixed sized arrays? 2010-10-29T12:43:19 [planet][time][player] 2010-10-29T12:43:40 there you go 2010-10-29T12:43:54 if that overflows, it fills into your struct (that appears unused) 2010-10-29T12:44:06 maybe you're off by one or something 2010-10-29T12:44:18 hmm, i'll have a poke around 2010-10-29T12:44:29 its usually the case anyway 2010-10-29T12:44:36 if you have other weird memory bugs, it could happen 2010-10-29T12:44:58 you can use valgrind to debug that kind of shit 2010-10-29T12:45:00 sometimes 2010-10-29T12:45:23 but raedwulf analysis sounds good to me 2010-10-29T12:46:00 also you can use at() instead of operator[], that makes sure you don't overflow 2010-10-29T12:46:29 yea, valgrind has a tendency to only deal with allocations though - if its fixed sized it might not find it 2010-10-29T12:46:50 i.e. whether the memory is allocated on the stack or on the heap 2010-10-29T12:46:59 heap? (maybe im wrong here) 2010-10-29T12:47:15 *** p4p4 has joined #aichallenge 2010-10-29T12:47:29 just googled, yea heap is the term used by system memory 2010-10-29T12:47:33 for* 2010-10-29T12:47:35 isn't the content of a vector always on the heap? 2010-10-29T12:47:52 the vector itself might be on the stack but it contains a vector to the heap 2010-10-29T12:47:54 im not sure, i think it depends on implementation 2010-10-29T12:48:00 but i might be completely wrong 2010-10-29T12:48:01 yea 2010-10-29T12:48:23 if he uses a "new" i think its always on heap 2010-10-29T12:48:53 isn't that what vector does internally? 2010-10-29T12:50:28 *** Bobng has quit IRC (Ping timeout: 255 seconds) 2010-10-29T12:50:31 *** nigg_ has joined #aichallenge 2010-10-29T12:51:45 not sure 2010-10-29T12:51:47 hehe 2010-10-29T12:52:08 that's why i use arrays 2010-10-29T12:52:11 yes vector does new internally 2010-10-29T12:52:22 vector op [] should do bounds checking 2010-10-29T12:52:51 maybe he's using simple arrays 2010-10-29T12:52:53 i'm pretty sure it doesn't 2010-10-29T12:52:59 at() does 2010-10-29T12:53:09 at least in the implementation i use 2010-10-29T12:53:50 and that causes nasty errors because depending on your compile flags, vector pads its contents, so you can go out of bounds without segfaulting 2010-10-29T12:54:28 *** irchs has quit IRC (Quit: irchs) 2010-10-29T12:55:10 hm MSVC's implementation has a #defined section in that operator that checks 2010-10-29T12:55:36 it's on in debug mode by default 2010-10-29T12:55:49 ah ok 2010-10-29T12:55:54 sounds right 2010-10-29T12:56:00 *** dmj111 has quit IRC (Read error: Connection reset by peer) 2010-10-29T12:56:07 i think i found the problem, strange bug, thanks 2010-10-29T12:56:08 I would expect it to be off in release mode 2010-10-29T12:56:17 *** dmj111 has joined #aichallenge 2010-10-29T12:56:29 i was potentially checking the "last" element of a vector that could be 0 2010-10-29T12:59:29 *** nigg_ has left #aichallenge ("Leaving") 2010-10-29T13:00:06 *** Bobng has joined #aichallenge 2010-10-29T13:04:11 *** janzert has joined #aichallenge 2010-10-29T13:04:42 Yay, after 2 days I have power back 2010-10-29T13:05:05 huzzah! 2010-10-29T13:06:11 now to try and catch up on what has happened. I see unfortunately no one took up running instances :( 2010-10-29T13:06:52 *** jaspervdj has quit IRC (Quit: NEVER GONNA GIVE YOU UP NEVER GONNA LET YOU DOWN) 2010-10-29T13:07:07 *** Bobng has quit IRC (Ping timeout: 255 seconds) 2010-10-29T13:11:54 running instances? 2010-10-29T13:12:36 Top 10 players: GreenTea(3882), bocsimacko(3822), rsergio(3777), felixcoto(3747), rs5147(3739), dmj111(3709), Accoun(3687), shangas(3671), protocolocon(3644), Hazard(3629) 2010-10-29T13:14:40 *** dmj111 has quit IRC (Read error: Connection reset by peer) 2010-10-29T13:14:56 *** dmj111 has joined #aichallenge 2010-10-29T13:15:36 *** bhasker has joined #aichallenge 2010-10-29T13:16:53 is jeff seriously changing the turns and mindistance to be completely different 2010-10-29T13:17:12 question: what is x going to be 2010-10-29T13:17:16 answer: !x 2010-10-29T13:17:36 turn limit change would be good 2010-10-29T13:17:53 min distance is currently 2, that's what they're currently planning to fix it at 2010-10-29T13:18:56 he completely reversed back to 200 later 2010-10-29T13:19:05 i don't know why more turns would be helpful i only go to 200 with stalemates 2010-10-29T13:19:44 *** bhasker has quit IRC (Ping timeout: 240 seconds) 2010-10-29T13:21:21 if anything game could be a bit shorter 2010-10-29T13:21:28 to allow for more games 2010-10-29T13:21:43 *** Olathe has quit IRC (Ping timeout: 276 seconds) 2010-10-29T13:23:42 reading the google group makes me even more confused as to who the hell is running this thing 2010-10-29T13:23:54 luckily i already switched to c++ so it doesn't matter 2010-10-29T13:24:54 mega1 thanks for your bot it motivated me to do the forecasting i had implemented but never used 2010-10-29T13:25:23 demerzel_: what do you forecast? 2010-10-29T13:25:42 well i have like 4 versions now 2010-10-29T13:25:52 but generally speaking i can tell you from current state any planet any turn's pop 2010-10-29T13:26:05 which i generally only used for defense purposes 2010-10-29T13:26:13 i also calc maximum fleet possibility to any planet on any turn 2010-10-29T13:26:27 which i need to figure out how not to always use ;) 2010-10-29T13:26:44 you are using an order buffer though which i implemented and never used 2010-10-29T13:27:11 *** bhasker has joined #aichallenge 2010-10-29T13:27:29 it seems you do linear expansion but i'm not sure if you're using distance or a minimax 'if i own this on turn 3 i can own that on turn 9' 2010-10-29T13:27:34 can't figure out your pattern haha 2010-10-29T13:30:10 janzert! You're back! 2010-10-29T13:30:38 yes! :) 2010-10-29T13:31:16 mega1: i'm coming after you sir 2010-10-29T13:31:17 did you see the script I wrote for you? 2010-10-29T13:31:24 mega1: i am currently trying to abstract all parts of the actual ships involved and looking at moves as solely capture/loss 2010-10-29T13:31:26 I see you got golang integrated into the setup script. Still catching up on a bunch of stuff but I'll look at it as soon as I can 2010-10-29T13:31:30 which is incredibly hard 2010-10-29T13:31:32 antimatroid: have you found so many bugs? 2010-10-29T13:31:37 yep 2010-10-29T13:31:53 and i have a new implementation to try soonish 2010-10-29T13:31:56 possibly not tonight though 2010-10-29T13:32:02 that's where most of my improvements came from, too 2010-10-29T13:32:13 i found soo many bugs fundamentally :P 2010-10-29T13:32:16 what is? 2010-10-29T13:32:21 bugs? 2010-10-29T13:32:50 so bad in fact i was potentially sending back moves with negative fleets that would then be ignored 2010-10-29T13:33:01 janzert, if you have any troubles with it let me know... 2010-10-29T13:36:01 don't worry I will :) 2010-10-29T13:37:38 *** hellman has quit IRC (Remote host closed the connection) 2010-10-29T13:38:09 hahaha: http://72.44.46.68/canvas?game_id=766958 2010-10-29T13:38:38 that game is a draw... 2010-10-29T13:39:13 both timed out? 2010-10-29T13:39:52 JamesMG, yours likes remote +1 planets almost as much as mine... 2010-10-29T13:39:59 yeah janzert 2010-10-29T13:40:07 *** bhasker_ has joined #aichallenge 2010-10-29T13:40:38 lavalamp__: yeah, I found some bugs in some of my methods while working on my other bot I need to move back to that one 2010-10-29T13:41:35 *** bhasker has quit IRC (Ping timeout: 252 seconds) 2010-10-29T13:44:02 *** Frontier has joined #aichallenge 2010-10-29T13:44:33 *** foucist has left #aichallenge 2010-10-29T13:46:56 *** bhasker_ has quit IRC (Quit: bhasker_) 2010-10-29T13:51:52 *** Florian has quit IRC (Remote host closed the connection) 2010-10-29T13:58:36 I like bugs that only appear when I'm losing 2010-10-29T13:59:45 they're better than the ones that crash your bot right before you win 2010-10-29T14:02:12 *** perestrelka_ has joined #aichallenge 2010-10-29T14:02:32 *** perestrelka has quit IRC (Ping timeout: 250 seconds) 2010-10-29T14:02:36 *** perestrelka_ is now known as perestrelka 2010-10-29T14:03:56 http://72.44.46.68/canvas?game_id=767741 2010-10-29T14:04:23 test failed 2010-10-29T14:04:40 harsh 2010-10-29T14:05:20 man i can't wait to get a working bot going 2010-10-29T14:05:25 map844 2010-10-29T14:05:32 its tempting to shortcut but if i do i'll never get these lofty goals implemented later 2010-10-29T14:05:37 that is so terrible 2010-10-29T14:05:41 they are too close 2010-10-29T14:05:57 *** dmj111 has quit IRC (Read error: Connection reset by peer) 2010-10-29T14:05:59 those maps should be replaced with maps from the corrected generator 2010-10-29T14:06:09 there's 3 levels of defense, those are the easiest to weed out people with 2/3 2010-10-29T14:06:21 *** dmj111 has joined #aichallenge 2010-10-29T14:06:22 Zannick: clearly they're not too trivial for everyone yet 2010-10-29T14:06:55 overlapping planets? 2010-10-29T14:07:03 planets with distance 1? 2010-10-29T14:07:08 oh, you want a better visualiser? 2010-10-29T14:07:16 oh meh, i want planets with distance one 2010-10-29T14:07:17 that was a bug that was fixed 2010-10-29T14:07:21 i don't know why people don't 2010-10-29T14:07:26 are they 1? 2010-10-29T14:07:33 whats a better visualizer 2010-10-29T14:07:40 it was a matter of the generator not applying the too_close check to the starting planets 2010-10-29T14:07:46 one that displays the planets not overlapping 2010-10-29T14:07:48 planets can be 1 apart 2010-10-29T14:08:15 but unfortunately you can't do that since with the old code the starting planets could be arbitrarily close 2010-10-29T14:08:54 *** narnach has quit IRC (Quit: Leaving.) 2010-10-29T14:08:55 have we got any maps yet that aren't symmetric 50 50 2010-10-29T14:09:08 like split it into thirds, and have a player start on one of the thirds 2010-10-29T14:09:10 they are generated that way 2010-10-29T14:09:16 with 0 being in the dead center 2010-10-29T14:09:22 that's boring 2010-10-29T14:09:26 i guess 2010-10-29T14:09:31 other maps would much more favour expansion 2010-10-29T14:09:34 i think 2010-10-29T14:09:44 really all it is now is superior expansion at start 2010-10-29T14:09:45 depends how good the planets are at a distance 2010-10-29T14:09:55 but there is a max best way to expand and if the other guy does the same then it's state keeping 2010-10-29T14:10:06 no it's not, my bot does half decently and it's starting move always sucks 2010-10-29T14:10:08 then if he keeps state as well as you then it's just sitting and waiting till turn 200 2010-10-29T14:10:47 plugging in all my billions of state shit in c++ is taking forever 2010-10-29T14:10:55 implementing the starting bot was like a couple hours 2010-10-29T14:12:37 Top 10 players: GreenTea(3905), bocsimacko(3846), rsergio(3801), felixcoto(3788), rs5147(3742), dmj111(3726), Accoun(3711), shangas(3704), Hazard(3677), protocolocon(3666) 2010-10-29T14:13:00 demerzel_: I'm not sure there is really a single optimal way to start out 2010-10-29T14:13:10 because of RageBot style openings 2010-10-29T14:13:11 there is that's why mega1 does so well 2010-10-29T14:13:34 eh? 2010-10-29T14:13:36 well one of several reasons but a major one 2010-10-29T14:13:40 you expand early extremely well 2010-10-29T14:13:54 cleanly and seems to have a covering check 2010-10-29T14:14:22 ah, I thought that's actually a weak spot 2010-10-29T14:14:35 no because your expansion targets are covered 2010-10-29T14:14:55 or i'm not sure exactly what the logic is but you don't just send to the closest best payback/cost:value 2010-10-29T14:15:10 which works because then you're never in a position where you're overextended 2010-10-29T14:15:51 and I got butchered when risk must be taken 2010-10-29T14:16:08 yea it's tough 2010-10-29T14:16:20 a true minimax bot would shut down if you didnt' at least match enemy expansion 2010-10-29T14:16:42 i had one that went all 200-ply and wouldn't do jack if i wasn't at least 1 ahead because it saw all moves as a risk 2010-10-29T14:17:37 i'm on like recode #5 and still playing with algorithms :( 2010-10-29T14:19:18 how on earth did you go 200 ply 2010-10-29T14:19:29 local, no timeout? 2010-10-29T14:19:32 i am an optimization freak 2010-10-29T14:19:35 no 2010-10-29T14:19:51 my first contest bot was 200-ply 2010-10-29T14:20:09 and considered every source->dest from 1 ship to how many it had before every move 2010-10-29T14:20:13 wow thats very impressive.... how many nodes per ply? 2010-10-29T14:20:22 in C# no less 2010-10-29T14:20:30 wow 2010-10-29T14:20:41 you can do a lot if you optimize 2010-10-29T14:20:46 people underestimate 2010-10-29T14:20:49 I'm having a hard time believing that 2010-10-29T14:20:59 you evaluated 9.41x10^23 possible moves in under 1 sec 2010-10-29T14:21:15 there aren't that many 2010-10-29T14:21:18 more 2010-10-29T14:21:20 yes there are 2010-10-29T14:21:26 even if there is only a single planet 2010-10-29T14:21:31 100 ships 2010-10-29T14:21:39 100 ships, 23 planets, 122 choose 22 2010-10-29T14:21:42 its 100^200 or so 2010-10-29T14:21:47 ah well what i did was set up a separate class 2010-10-29T14:21:53 that kept a simulated world 2010-10-29T14:21:55 and that's just on the first turn 2010-10-29T14:22:09 it updated source/dest turns for scenario x 2010-10-29T14:22:14 then reset it to 'actual' state 2010-10-29T14:22:17 kept score 2010-10-29T14:22:32 i think it might have just been exiting early every time 2010-10-29T14:22:36 no 2010-10-29T14:22:44 it was close 2010-10-29T14:22:56 my slowest bot by far but it made 1s it was like ~150 2010-10-29T14:23:12 i would like to see this after the contest 2010-10-29T14:23:18 what kind of computer do you have and where can I get one? 2010-10-29T14:23:23 if you really got that far 2010-10-29T14:23:24 on the contest server :P 2010-10-29T14:24:00 you're saying you try every combination of ships from every source planet to every dest planet? 2010-10-29T14:24:20 i don't any more 2010-10-29T14:24:26 that was my first attempt at minmax 2010-10-29T14:24:30 but you did? 2010-10-29T14:24:46 every combination of ships from my planets to all planets yes 2010-10-29T14:24:52 how many nodes were there at ply 200? :D 2010-10-29T14:25:03 more than atoms in the universe lol 2010-10-29T14:25:16 I will bet you money that you had a bug that causes it to not evaluate every combination 2010-10-29T14:25:17 it didn't keep nodes it kept best score 2010-10-29T14:25:27 purely score-based eval 2010-10-29T14:25:42 any pruning? alpha-beta/whatnot? 2010-10-29T14:25:51 not at first, eventually yes 2010-10-29T14:26:13 this game has a much higher branching factor than go and it can't be done for go 2010-10-29T14:26:14 i still have the code somewhere i'd post it idc i've recoded twice and am in a completely diff language now 2010-10-29T14:26:53 basically it had a Holodeck class that allowed me to simulate as many moves as i wanted then record the score, the nreset it 2010-10-29T14:27:10 if it got a better score it would hold the source/target/fleetsize 2010-10-29T14:27:24 are you sure you weren't doing monte carlo? that doesn't sound like minimax 2010-10-29T14:27:28 that is one way one could do minimax, but it is normally not very efficient 2010-10-29T14:27:35 it's not efficient at all lol 2010-10-29T14:27:44 but it ran in under a second 2010-10-29T14:27:46 only reason it worked is i am incredibly anal about performance 2010-10-29T14:27:53 so you get to 200ply on a single branch 2010-10-29T14:28:01 or a couple 2010-10-29T14:28:18 one branch per move, do x then y then z 2010-10-29T14:28:24 add totals 2010-10-29T14:28:49 it's a terribly shitty system! all i am saying is people underestimate what even a shitty modern computer can do was my point 2010-10-29T14:28:55 yeah, if you can't find your source and post it after the contest, we will probably not believe you 2010-10-29T14:29:01 lol 2010-10-29T14:29:24 since what you're claiming is to have evaluated every possible state 2010-10-29T14:29:25 sorry, I don't believe you were exhaustively searching the space 2010-10-29T14:29:31 i probably wasn't 2010-10-29T14:29:41 if you succeeeded your bot would have owned everyone 2010-10-29T14:29:43 it still did a fuckton of calculations in under a second 2010-10-29T14:30:26 demerzel_: maybe if you port it to c++ you can solve the entire game. ;) 2010-10-29T14:30:34 haha i am working on it 2010-10-29T14:30:47 c++ is strangely freeing 2010-10-29T14:30:56 i don't have to worry about whether the compiler will do this or how to init it 2010-10-29T14:31:02 i just /know/ 2010-10-29T14:31:09 but your point is probably true anyway... 2010-10-29T14:31:27 a perfect minimax wouldn't do anything on move 0 2010-10-29T14:31:43 huh? 2010-10-29T14:31:49 i doubt that 2010-10-29T14:31:52 no it would 2010-10-29T14:31:58 it just wouldn't do anything if it was losing 2010-10-29T14:32:07 huh? 2010-10-29T14:32:09 or rather - lock down 2010-10-29T14:32:16 i doubt that, too 2010-10-29T14:32:20 minimax selects the best choice even if it is negative 2010-10-29T14:32:29 yes the choice is inactivity 2010-10-29T14:32:40 if your state has 'how much enemy can get to me' 2010-10-29T14:32:45 and also 'how much i can get there' 2010-10-29T14:32:49 and yours is less than the enemy 2010-10-29T14:32:52 why? surely best choice when you are losing is to make as much trouble as possible and hope for a mistake 2010-10-29T14:32:53 then it won't fire 2010-10-29T14:33:04 apply that to my planets 2010-10-29T14:33:07 the first one is false because there are maps (eg map1) that have starter planets at distance > 20 so inactivity is strictly worse than taking any neutral 2010-10-29T14:33:09 every planet is under risk of loss 2010-10-29T14:33:17 welll excepting weird maps ofc 2010-10-29T14:33:28 lavalamp__: that's not minimax :) 2010-10-29T14:33:38 good point 2010-10-29T14:33:39 no that's a flaw of minmax actually 2010-10-29T14:33:51 it assumes the enemy will make the best move every time 2010-10-29T14:34:06 you could redesign some part of the scoring to be better when there are less losing states in the tree 2010-10-29T14:34:12 my current bot attacks every time with exactly how much is needed to not allow the enemy ot reinforce 2010-10-29T14:34:20 that's a very slight improvement/hope for mistake 2010-10-29T14:34:21 because it can't make a decision as to whether it's necessary or not 2010-10-29T14:34:28 Ah hm maybe you guys are right 2010-10-29T14:34:38 hope this complexity pays off in the end 2010-10-29T14:35:02 eh this is fun to me lol 2010-10-29T14:35:28 well I do recursive payoff tables and I don't think my bot has that problem... hm 2010-10-29T14:35:55 your payoff tables are subjective 2010-10-29T14:36:00 or your state is incomplete 2010-10-29T14:37:07 I give each node two scores (mine, theirs)-- I make it maximize its own score 2010-10-29T14:37:43 I'm not sure if that counts as subjective but my search is definitely not exhaustive 2010-10-29T14:37:50 *** dmj111 has quit IRC (Read error: Connection reset by peer) 2010-10-29T14:38:08 *** dmj111 has joined #aichallenge 2010-10-29T14:39:45 plain minimax is very bad at "hoping for a mistake" 2010-10-29T14:40:20 oh that was basically said already 2010-10-29T14:40:30 * jmcarthur should finish reading a conversation before piping in 2010-10-29T14:40:34 heh 2010-10-29T14:42:01 yeah I'm mistaken, I keep getting confused because I'm not doing straight minimax 2010-10-29T14:42:33 that's actually all i know i am trying to figure out some other concepts 2010-10-29T14:42:40 either that or i guess i can try to power through it with c++ now lol 2010-10-29T14:42:49 hm... in fact I should add code to choose more randomly when behind... 2010-10-29T14:46:59 lavalamp__: does your scoring function have a component to prioritize targetting the enemy player? 2010-10-29T14:47:18 [11:40] I give each node two scores (mine, theirs)-- I make it maximize its own score 2010-10-29T14:47:21 that makes it sound like no? 2010-10-29T14:47:57 ummmm 2010-10-29T14:48:12 feel free not to answer 2010-10-29T14:48:14 *** perror has quit IRC (Quit: Bye all !) 2010-10-29T14:48:21 I'm not quite sure 2010-10-29T14:48:28 oh, haha 2010-10-29T14:48:31 traditional minimax uses one score 2010-10-29T14:48:31 it would 2010-10-29T14:48:36 mine-theirs, for instance 2010-10-29T14:48:41 so if you did that, it would 2010-10-29T14:48:43 i just use negative for enemy 2010-10-29T14:48:49 yeah Zannick, mine does that 2010-10-29T14:48:57 but that would just be the same thing really, 0 vs 5 = -5 2010-10-29T14:48:59 mine-thiers > current best mine-thiers 2010-10-29T14:49:13 Actually I maximize mine-theres, but if that is 0 I maximize mine 2010-10-29T14:49:26 because some 0's are better than others 2010-10-29T14:50:12 I started off with just one score for both but it was too hard to add modifiers onto it 2010-10-29T14:50:46 and I couldn't make it tell the difference between we both expand and we both do nothing 2010-10-29T14:51:20 plus this way is generalizable to > 2 player games 2010-10-29T14:51:30 not that that matters in any way 2010-10-29T14:51:47 I want to try making it adjustable players after the contest 2010-10-29T14:53:49 *** RainCT has joined #aichallenge 2010-10-29T14:55:19 *** n9thbit has quit IRC (Quit: n9thbit) 2010-10-29T14:58:26 unfortunately multiple players would require an n-dimensional payoff matrix 2010-10-29T14:58:58 as if my node count weren't already woefully inadequate 2010-10-29T14:59:25 yeah, i have the struct for nd payoff tensors 2010-10-29T14:59:41 but the size of them explodes as the dimensions blow out, both with strategies and number of players 2010-10-29T15:05:05 *** Accoun_ has quit IRC () 2010-10-29T15:11:10 *** AlliedEnvy has quit IRC (Ping timeout: 265 seconds) 2010-10-29T15:12:20 *** Frontier has quit IRC (Ping timeout: 252 seconds) 2010-10-29T15:12:37 Top 10 players: GreenTea(3900), bocsimacko(3872), felixcoto(3834), rsergio(3787), rs5147(3754), dmj111(3748), Accoun(3745), Hazard(3713), shangas(3709), protocolocon(3684) 2010-10-29T15:15:52 whoa, mega1 isn't in first 2010-10-29T15:16:20 who is mega1 on official rankings? 2010-10-29T15:16:26 bocsimacko 2010-10-29T15:16:30 ahh 2010-10-29T15:16:46 he's untouchable on the TCP server 2010-10-29T15:16:55 who is GreenTea? 2010-10-29T15:17:04 *** Accoun has joined #aichallenge 2010-10-29T15:17:56 *** AlliedEnvy has joined #aichallenge 2010-10-29T15:24:59 hmmm, dabino is using mixed strategy nash equilibria? 2010-10-29T15:26:17 what makes you think that 2010-10-29T15:26:18 ? 2010-10-29T15:28:10 look at their name on tcp 2010-10-29T15:28:18 dabino_msne 2010-10-29T15:29:12 ah :) wouldn't have thought of that 2010-10-29T15:31:09 doesn't everyone? ;+ 2010-10-29T15:31:14 er, ;) 2010-10-29T15:31:53 with all the talk about payoff matrices here, maybe :) 2010-10-29T15:32:25 that's probably just a handful of us doing too much talking :) 2010-10-29T15:32:26 *** dmj111 has quit IRC (Read error: Connection reset by peer) 2010-10-29T15:32:41 *** dmj111 has joined #aichallenge 2010-10-29T15:34:05 my current bot is now at a crawling pace 2010-10-29T15:35:49 I went from 20k nps to like 1k 2010-10-29T15:48:24 C++ compiles so slow 2010-10-29T15:51:17 *** rebelxt has joined #aichallenge 2010-10-29T15:53:52 *** dmj111 has quit IRC (Ping timeout: 245 seconds) 2010-10-29T15:54:11 *** dmj111 has joined #aichallenge 2010-10-29T15:55:08 *** n9thbit has joined #aichallenge 2010-10-29T15:56:19 but it compiles so right 2010-10-29T15:56:22 *if you code it right 2010-10-29T16:01:31 The build I just finished took 9:30 2010-10-29T16:03:33 *** wvdschel has joined #aichallenge 2010-10-29T16:03:33 *** wvdschel has joined #aichallenge 2010-10-29T16:03:48 jesus what is in your build 2010-10-29T16:04:04 let me guess, boost? 2010-10-29T16:04:04 mixed strategy to me is just fancytalk for 'heuristics' unless it's a true strat profile which i doubt 2010-10-29T16:04:19 yeah we have part of boost in it 2010-10-29T16:04:30 that template magic... 2010-10-29T16:04:35 This is my day job, not my bot, btw 2010-10-29T16:04:42 *** Naktibalda_ has joined #aichallenge 2010-10-29T16:04:42 oh hahaha 2010-10-29T16:04:45 and that number is with a friggin compile farm 2010-10-29T16:05:03 boost is crap just install visual studio and give up 2010-10-29T16:05:11 ... 2010-10-29T16:05:12 oh microsoft which functions should i use 2010-10-29T16:05:25 scanf is deprecated? yes sir let me use yours instead 2010-10-29T16:05:34 hahaha 2010-10-29T16:05:39 uhh 2010-10-29T16:06:08 sadly that kind of dev exists 2010-10-29T16:06:34 that thread with the dude that was asking what to do with .cc files and needed a walkthrough of adding them to a vs project blew my mind 2010-10-29T16:06:58 there have been a few c++ threads that have astonished me 2010-10-29T16:07:19 i guess people are using this to learn programming/a new language 2010-10-29T16:07:30 i barely know c++ i think i last used it when i was like 15 2010-10-29T16:07:55 amazingly C# is close enough to let me remember most of it 2010-10-29T16:08:11 initialization of arrays and shit cost me a few hours of google time but otherwise it's pretty smooth - fits my dev style perfectly 2010-10-29T16:08:37 the fact that you need to initialize them cost you hours? 2010-10-29T16:08:43 or the syntax of how to do it 2010-10-29T16:08:53 well i always research everything beyond just syntax 2010-10-29T16:09:03 I always have to google for one thing, the syntax for pointers to member functions 2010-10-29T16:09:14 I can never remember 2010-10-29T16:09:15 in c++? 2010-10-29T16:09:19 init lists then trying to init arrays in them etc 2010-10-29T16:09:27 why would you want a pointer to member functions? that sounds dangerous 2010-10-29T16:09:36 lol it's not useful very often 2010-10-29T16:09:56 actually, not true 2010-10-29T16:09:58 anonymous functions? 2010-10-29T16:10:00 i do love the shit out of pointers 2010-10-29T16:10:08 std::for_each is an example of something that can use that 2010-10-29T16:10:14 *** rebelxt has left #aichallenge 2010-10-29T16:10:22 i don't even know why i'm using c++ honestly i have not a single std:: anything 2010-10-29T16:10:27 along with std::memf or whatever it is 2010-10-29T16:10:34 bind something or other 2010-10-29T16:10:35 i guess cuz i don't know the c syntax for class/structs or if they're even possible 2010-10-29T16:10:51 structs are, classes no 2010-10-29T16:11:01 ah there we go then 2010-10-29T16:11:12 i still oop but i much prefer the control of c standard rather than c++ 2010-10-29T16:11:43 well nobody stops you from doing things c style in c++ 2010-10-29T16:11:57 its a subset 2010-10-29T16:12:00 almost, i think 2010-10-29T16:12:07 yea except every goddamn time i google 'array c++' i have to sift through 20 posts about how C++ has vectors goddamnit they're so awesome why are you still using shitty arrays blahblah 2010-10-29T16:12:13 or god forbid c strings 2010-10-29T16:12:22 hahaha 2010-10-29T16:12:38 Top 10 players: GreenTea(3923), bocsimacko(3891), felixcoto(3886), rsergio(3810), rs5147(3775), dmj111(3770), Accoun(3764), Hazard(3761), shangas(3746), protocolocon(3693) 2010-10-29T16:12:39 ya c strings really suck... 2010-10-29T16:12:45 i like them :P 2010-10-29T16:12:55 i wish they were pascal strings but i still like them 2010-10-29T16:12:56 for optimization ok 2010-10-29T16:13:04 lol they're so hard to work with... 2010-10-29T16:13:04 but in a production environment... 2010-10-29T16:13:17 they're not that bad 2010-10-29T16:13:25 i just wish they had a standard pascal option 2010-10-29T16:13:31 if they carried length they'd be great 2010-10-29T16:13:52 hehe isnt that really what a std::string is 2010-10-29T16:13:57 meh all I know is I'm at least twice as fast manipulating go strings 2010-10-29T16:14:00 a c string that knows its length 2010-10-29T16:14:09 and does all the boilerplate allocation for you 2010-10-29T16:14:15 std::string isn't too bad 2010-10-29T16:14:35 lavalamp__: you use go? 2010-10-29T16:14:35 cool 2010-10-29T16:14:39 havent tried it yet 2010-10-29T16:14:45 yeah my bot is in go 2010-10-29T16:14:50 no i would use std::string 2010-10-29T16:14:59 no wonder you complain about c++ compile times 2010-10-29T16:15:00 but i don't believe streams are better than direct mem allocation 2010-10-29T16:15:02 because theyre not 2010-10-29T16:15:03 hopefully janzert will accept my script for installing go on the server 2010-10-29T16:15:27 to be fair, I don't have any million plus LOC go apps to compare, but yeah... 2010-10-29T16:15:29 my dev career has gone the opposite way of most 2010-10-29T16:15:36 start in easy mode and demand hard mode 2010-10-29T16:15:42 *** wvdschel has quit IRC (Quit: Leaving) 2010-10-29T16:15:47 i am sure if i started with c i'd love all the neato shit c# has 2010-10-29T16:15:58 *** wvdschel has joined #aichallenge 2010-10-29T16:15:59 nah 2010-10-29T16:16:05 but i view it as an obstacle to my optimization 2010-10-29T16:17:08 also holy shit i love blah::mystruct = {0} 2010-10-29T16:17:20 why that isn't in all the 'convenient' langs is beyond me 2010-10-29T16:17:32 go has 2010-10-29T16:17:39 x := mystruct{} 2010-10-29T16:18:17 i was amazed the first time i used it 2010-10-29T16:18:26 my struct with 4 arrays was wholly initialized arrays and all to 0 2010-10-29T16:18:34 and no assignment 2010-10-29T16:18:53 oh, go does that when you declare any var 2010-10-29T16:19:23 I thought most langs zeroed stuff for you 2010-10-29T16:20:01 well most langs don't have a 2 step process between dec and init 2010-10-29T16:23:24 demerzel_: hey 2010-10-29T16:24:04 hey 2010-10-29T16:24:17 *** Naktibalda_ has quit IRC (Ping timeout: 272 seconds) 2010-10-29T16:24:30 someone reported you for this post: http://ai-contest.com/forum/viewtopic.php?f=18&t=1025 2010-10-29T16:24:39 i understand it's a noob question, but you could be a little nicer 2010-10-29T16:24:49 haha 2010-10-29T16:24:51 come on 2010-10-29T16:24:54 yeah, lol 2010-10-29T16:25:02 it would take less effort to google it than create a post 2010-10-29T16:25:17 *** Naktibalda_ has joined #aichallenge 2010-10-29T16:25:59 i hardly ever get to use lmgtfy.com any more thought it was appropriate ;) 2010-10-29T16:27:07 lol 2010-10-29T16:29:01 i figured if anything you'd be more pissed at me for my incessant bitching about c# 2010-10-29T16:29:09 *** zoo has joined #aichallenge 2010-10-29T16:29:15 which i have ceased since recognizing c++ as the superior language ;) 2010-10-29T16:32:56 I won't be able to work on my bot tomorrow :( 2010-10-29T16:34:02 *** dmj111 has quit IRC (Read error: Connection reset by peer) 2010-10-29T16:34:21 *** dmj111 has joined #aichallenge 2010-10-29T16:35:49 demerzel_: that too 2010-10-29T16:37:03 @where visualizer 2010-10-29T16:37:03 amstan: visualizer = (#1) http://github.com/DanielVF/Planet-Wars-Canvas-Visualizer, or (#2) http://csclub.uwaterloo.ca/~amstan/aichallenge/canvas-visualizer/ 2010-10-29T16:38:01 sorry! tbh tho if i did continue in C# and it kept screwing up i'd probably be pretty pissed at spending a month on something that has no chance in the finals 2010-10-29T16:38:22 speaking of danielVF does he cache the old games somewhere? they are cleaned ridiculously quickly 2010-10-29T16:38:24 that's what happened last year with java 2010-10-29T16:38:38 no shit haha 2010-10-29T16:38:48 well i mean i did code with the fact that it's not binary in mind ie optimization 2010-10-29T16:38:49 if you pick these languages you have to keep in mind that there's problems with implementations 2010-10-29T16:38:52 and they might not always work 2010-10-29T16:38:54 but still 2010-10-29T16:38:57 it's a risk you have to take 2010-10-29T16:39:01 i can't help it if the whole thing don't work 2010-10-29T16:39:04 *** bhasker has joined #aichallenge 2010-10-29T16:39:40 C# was not designed to be used like this. C# was made for visual studio, to make apps in windows. Then they ported it to linux to gain more popularity. 2010-10-29T16:39:56 * janzert thinks demerzel_ is confusing danielvf and dhartmei 2010-10-29T16:40:00 yea i am 2010-10-29T16:40:04 *** Arthur_ has quit IRC (Ping timeout: 240 seconds) 2010-10-29T16:40:13 does he cache the games somewhere 2010-10-29T16:40:21 not that I know of 2010-10-29T16:40:26 yea well then don't offer them :P 2010-10-29T16:40:38 *** p4p4 has quit IRC (Quit: ChatZilla 0.9.84 [SeaMonkey 2.0a3/20090223135443]) 2010-10-29T16:41:01 if i say come inside i have 101 flavors of ice cream and really i only have 6 and the rest is just colored milk i'm pretty sure you wouldn't just say 'welp fridges break sometimes!' 2010-10-29T16:41:09 it worked until people started using it :P 2010-10-29T16:41:35 esp if you'd been waiting in line for a month planning your order haha 2010-10-29T16:41:40 since it was a resource leak that took a while to show up 2010-10-29T16:41:47 yea i'm sure it was 2010-10-29T16:41:56 did you see my post on the script to make a parallel ? 2010-10-29T16:42:15 http://72.44.46.68/canvas?game_id=769108 2010-10-29T16:42:18 I like that map a lot 2010-10-29T16:42:19 not that it matters to me any more but might be useful 2010-10-29T16:42:27 parallel question marks? 2010-10-29T16:42:28 not so much the way my bot handled it... 2010-10-29T16:42:49 no a parallel mono deployment for the stable version 2010-10-29T16:42:52 on ubuntu 2010-10-29T16:43:05 i say parallel because you would keep the packaged one too 2010-10-29T16:43:14 since you didn't want to go outside package 2010-10-29T16:43:29 ahh, I think I saw it but I don't want to build from scratch if I can help it and most people are happy with 2.4 it seems 2010-10-29T16:44:00 you don't have to build it from scratch a guy wrote a script to do it for you practically 2010-10-29T16:44:06 I wouldn't have any *need* to keep the package one if I had to use the one from scratch 2010-10-29T16:44:26 and badgerports has tested packages 2010-10-29T16:44:42 just not for 2.6 and ubuntu hardy :) 2010-10-29T16:44:49 *shrug* 2010-10-29T16:45:04 doesn't matter to me any more but that's a pretty well-known script ie it's not just some dude on a blog i know it's been used a lot 2010-10-29T16:45:11 personally I'm done with C#, if another admin wants to step in and do more it's up to them :/ 2010-10-29T16:45:46 thanks for validating my switch to c++ then haha 2010-10-29T16:45:46 but how much with hardy would be the question I think 2010-10-29T16:45:49 *** yasith has quit IRC (Ping timeout: 272 seconds) 2010-10-29T16:46:18 well like I said near the beginning of the week if it were just up to me I'd have yanked support for it altogether 2010-10-29T16:46:32 it was way more trouble than it was worth in my view 2010-10-29T16:46:51 you're using hardy? i thought it was just a 9.1 box 2010-10-29T16:46:58 I'll port to go if you get that working, so I can continue to be a pain in your ass janzert 2010-10-29T16:47:02 idk then 2010-10-29T16:47:11 we wish it were more modern :P 2010-10-29T16:47:12 JamesMG: hehe 2010-10-29T16:47:42 yank it then haha 2010-10-29T16:47:48 or start offering VB.NET 2010-10-29T16:47:52 can't wait for that flood 2010-10-29T16:48:11 it's working now so I ain't touchin' it :P 2010-10-29T16:48:31 are you C# MG? 2010-10-29T16:48:41 i don't ever look on the official site but i recognize your name from tcp 2010-10-29T16:48:56 yeah, I'm on C# 2010-10-29T16:49:08 i'm not sure what extension i have in firefox that overrides the "leaderboard" class but i can't ever see it at home 2010-10-29T16:49:11 originally did that to learn C#, oops. 2010-10-29T16:49:14 haha 2010-10-29T16:49:23 "oops" 2010-10-29T16:49:24 haha 2010-10-29T16:49:46 I haven't had a loss I didn't deserve since you "fixed" it janzert 2010-10-29T16:49:52 same 2010-10-29T16:49:56 i had like 9 or 10 in a row 2010-10-29T16:50:02 i'm actually <200 now with my busted ass bot 2010-10-29T16:50:08 9 or 10 crashes that is 2010-10-29T16:50:13 yeah, same here 2010-10-29T16:50:21 great 2010-10-29T16:50:38 I think it really is fixed or rather the work around is working 2010-10-29T16:50:45 hey quick q 2010-10-29T16:50:50 *** perestrelka has quit IRC (Read error: Connection reset by peer) 2010-10-29T16:51:00 does having a universal header file and multiple cc/cpp files mess up the gcc compiler 2010-10-29T16:51:09 i am using vs lol 2010-10-29T16:51:25 *** perestrelka has joined #aichallenge 2010-10-29T16:51:36 not positive what you mean, but if I understand you correctly it shouldn't 2010-10-29T16:51:43 not sure if it expects a 1:1 .h:.cpp or how it gathers files 2010-10-29T16:51:58 no 2010-10-29T16:52:09 cool ty 2010-10-29T16:52:10 the compiler basically doesn't care 2010-10-29T16:52:26 you want to know what #include does? 2010-10-29T16:52:28 in cpp .h files are inserted textually into the source file before compilation 2010-10-29T16:52:34 pastes the entire thing in 2010-10-29T16:52:52 yeah, the preprocessor does major hax 2010-10-29T16:52:56 hmm so a universal header file is bad? i thought that's what include guards prevented 2010-10-29T16:53:03 no, it isn't bad 2010-10-29T16:53:08 i just got tired of updating similar .h and ordering them precisely 2010-10-29T16:53:32 i have never used c++ at a job so i don't know standards at all lol 2010-10-29T16:53:36 as long as header files are only declarations, not definitions 2010-10-29T16:53:44 they'd probably fire me if i tried here haha 2010-10-29T16:54:14 declarations mean yes, this function name/variable is somewhere, so the code file can be compiled 2010-10-29T16:54:26 yea no i know what they are i just know very little about standards 2010-10-29T16:54:33 then the linker will paste things together so the binary has the referenced thing 2010-10-29T16:54:38 outside of init lists are good and don't be a dumbass with pointers 2010-10-29T16:54:55 i have a global debug.h that has a bunch of #defines 2010-10-29T16:55:22 do you ever use forward decs? i have never seen anyone use them ever i didn't evne know they were possible 2010-10-29T16:56:03 what do you mean? 2010-10-29T16:56:46 stub declarations basically 2010-10-29T16:56:52 my .h files have forward declarations, that's how that generally is done 2010-10-29T16:57:00 well no i meant in the actual cpp file 2010-10-29T16:57:17 i used to do that, but instead i just include the header itself 2010-10-29T16:57:26 ah ok yea i've always done that i just wondered 2010-10-29T16:58:09 sorry i am in a weird position with c++ - i know programming concepts and the language very well but have no idea on standards and nobody i know knows c++ not even the xkcd types lol 2010-10-29T16:58:30 haha, "xkcd types" 2010-10-29T16:58:44 i am not perfect on c++ though i use it at work, among others 2010-10-29T16:58:47 i prefer c 2010-10-29T16:59:26 i like c a lot it was my second language after the original BASIC 2010-10-29T16:59:37 i still prefer c strings and arrays and basically just use c++ for easier oop 2010-10-29T16:59:42 I C++ all day at work 2010-10-29T17:00:13 *** nullkuhl_ has joined #aichallenge 2010-10-29T17:00:29 i need to move to a real tech city i live in kansas city pretty much the only dev jobs are corp db ms shops 2010-10-29T17:00:48 does anyone here know good heuristics for the 8puzzle game ? 2010-10-29T17:00:58 link? 2010-10-29T17:03:15 holy crap i wish i understood this crap about backpropagation anns 2010-10-29T17:07:05 *** demerzel_ has quit IRC (Quit: Page closed) 2010-10-29T17:07:34 lavalamp__: http://www.8puzzle.com/ 2010-10-29T17:10:51 *** dmj111 has quit IRC (Read error: Connection reset by peer) 2010-10-29T17:11:13 *** dmj111 has joined #aichallenge 2010-10-29T17:11:24 nullkuh1_: no idea 2010-10-29T17:12:09 there's a trick involving multiplying each number by a fixed number 2010-10-29T17:12:21 like, a grid 1, -1, 1 or something. 2010-10-29T17:12:29 and then some values are impossible 2010-10-29T17:12:38 Top 10 players: GreenTea(3930), felixcoto(3905), bocsimacko(3897), rsergio(3798), Accoun(3794), rs5147(3784), dmj111(3767), Hazard(3758), shangas(3733), protocolocon(3703) 2010-10-29T17:13:44 i can't remember how to do it 2010-10-29T17:15:52 *** yasith has joined #aichallenge 2010-10-29T17:16:08 *** lavalamp__ has quit IRC (Quit: Page closed) 2010-10-29T17:16:48 *** delt0r has joined #aichallenge 2010-10-29T17:17:14 *** delt0r___ has quit IRC (Ping timeout: 240 seconds) 2010-10-29T17:30:57 *** dmj111 has quit IRC (Ping timeout: 245 seconds) 2010-10-29T17:31:53 *** dmj111 has joined #aichallenge 2010-10-29T17:33:25 *** dmj111 has left #aichallenge 2010-10-29T17:58:20 *** RainCT has quit IRC (Remote host closed the connection) 2010-10-29T18:02:03 *** TwoPixelGrid has quit IRC (Quit: Leaving) 2010-10-29T18:06:32 *** bhasker_ has joined #aichallenge 2010-10-29T18:08:39 *** yasith has quit IRC (Ping timeout: 265 seconds) 2010-10-29T18:09:22 *** bhasker has quit IRC (Ping timeout: 255 seconds) 2010-10-29T18:09:42 *** bhasker has joined #aichallenge 2010-10-29T18:10:57 *** Appleman1234 has joined #aichallenge 2010-10-29T18:11:53 *** amstan has quit IRC (Remote host closed the connection) 2010-10-29T18:12:39 Top 10 players: bocsimacko(3903), felixcoto(3894), GreenTea(3882), rsergio(3827), Accoun(3780), rs5147(3777), Hazard(3766), dmj111(3741), shangas(3740), joakim_mjardner(3740) 2010-10-29T18:13:06 *** amstan has joined #aichallenge 2010-10-29T18:13:06 *** ChanServ sets mode: +o amstan 2010-10-29T18:13:13 *** bhasker_ has quit IRC (Ping timeout: 264 seconds) 2010-10-29T18:16:42 *** JamesMG has quit IRC (Quit: Leaving.) 2010-10-29T18:21:52 *** Mathnerd314 has joined #aichallenge 2010-10-29T18:26:10 *** narnach has joined #aichallenge 2010-10-29T18:30:06 *** Azrathud has joined #aichallenge 2010-10-29T18:32:43 *** yasith has joined #aichallenge 2010-10-29T18:33:30 Does anyone know how wide the map is? 2010-10-29T18:33:42 I'm guessing 25 but I want to make sure 2010-10-29T18:34:05 doesn't matter; calculate it yourself. also see the map generator 2010-10-29T18:34:20 which might be made the official map maker 2010-10-29T18:34:22 *** wvdschel has quit IRC (Quit: Leaving) 2010-10-29T18:34:42 also look @ the forum which answered this already 2010-10-29T18:35:18 blargh, these irc logs are so hard to skim through for ideas 2010-10-29T18:35:57 Why doesn't it matter? Where on the forum? 2010-10-29T18:36:48 search for "map radius" 2010-10-29T18:37:38 planets are chosen within a square [-12,12) x [-12,12) 2010-10-29T18:37:51 for maps generated with the map generator 2010-10-29T18:38:12 ok, thanks 2010-10-29T18:52:30 *** McLeopold has joined #aichallenge 2010-10-29T18:53:55 paging jmcarthur... 2010-10-29T18:54:01 *** Cold-Phoenix has quit IRC (Ping timeout: 264 seconds) 2010-10-29T18:54:09 and anyone else interested in puzzels 2010-10-29T18:54:18 puzzles? 2010-10-29T18:54:20 oooh, puzzles? 2010-10-29T18:54:35 are you ready for this? 2010-10-29T18:55:22 depends on whether you're going to post it 2010-10-29T18:55:35 I'm trying to build up enthusiasm... 2010-10-29T18:55:51 heh 2010-10-29T18:56:04 I've got a rock / paper / sissors situation 2010-10-29T18:56:41 *** demerzel_ has joined #aichallenge 2010-10-29T18:56:41 *** Cold-Phoenix has joined #aichallenge 2010-10-29T19:00:24 a real one? 2010-10-29T19:01:05 ooh 2010-10-29T19:01:22 McLeopold: i wanna see it 2010-10-29T19:01:40 *** yasith has quit IRC (Ping timeout: 276 seconds) 2010-10-29T19:02:04 http://pastebin.com/Mr2PHG1G 2010-10-29T19:02:30 so you have 2 options, go for the close one, go for the furthur one or wait for a snipe 2010-10-29T19:02:55 close beats wait, wait beats far, far beats close 2010-10-29T19:03:34 is that a map with forethought 2010-10-29T19:03:35 ban 2010-10-29T19:03:55 you don't have enough ships to take both 2010-10-29T19:04:05 starting planets are 0 growth 2010-10-29T19:04:10 so you never will 2010-10-29T19:04:14 i would just paste that into my "bot" which is pure state atm but it would break it being <23 plaents 2010-10-29T19:04:39 0 growth starting planets is an interesting concept 2010-10-29T19:05:10 demerzel_: does a constant 23 planets gain you anything? 2010-10-29T19:05:31 it seems like you could just count them at the beginning and go with that 2010-10-29T19:05:44 not having to add every planet-based array (a lot) into a 'resizeItIfThere'sMoreOrLess' function in c++ no less 2010-10-29T19:05:59 i did do that for a long time in C# even it was a pain in teh ass 2010-10-29T19:06:15 in c++ it would be a nightmare because most of them are in static private structs 2010-10-29T19:06:33 i just allocate after reading in all the planets 2010-10-29T19:06:48 well i try to allocate as much as possible in the beginning 2010-10-29T19:06:51 2 seconds that is 2010-10-29T19:07:03 the initial planet list is 64, then everything is put into right-sized lists 2010-10-29T19:07:06 anyway 2010-10-29T19:07:21 lists? 2010-10-29T19:07:25 jmcarthur: comments? 2010-10-29T19:07:43 McLeopold: i'm gonna lay it out visually. gimme a few 2010-10-29T19:07:43 0 growth starting planets would make more interesting games 2010-10-29T19:07:52 since you'd have to choose a starting point instead of being given it 2010-10-29T19:08:21 how about 0 growth and 1000 ships. 2010-10-29T19:08:27 visual layout would be good 2010-10-29T19:08:34 too high i'd even go lower than 100 2010-10-29T19:08:43 i can't immediately recognize the distances here 2010-10-29T19:08:48 hold on 2010-10-29T19:08:51 enough to capture 1 of 4 (2 symmetrical planets) 2010-10-29T19:09:10 the stanford job fell through :( 2010-10-29T19:09:21 aw 2010-10-29T19:09:30 i am disappoint 2010-10-29T19:09:47 oh well. wasn't very high paying anyway 2010-10-29T19:09:52 just would have been fun 2010-10-29T19:10:56 maybe i should win this shit so some real dev company will relocate me to someplace better 2010-10-29T19:11:36 btw does anyone know of a code repository that isn't public only 2010-10-29T19:11:47 i am fine with publishing after but not really wanting a google search to find my source 2010-10-29T19:11:53 bitbucket allows private repos 2010-10-29T19:11:56 i am just dumping zips into mediafire atm 2010-10-29T19:12:01 hmm ty 2010-10-29T19:12:39 Top 10 players: bocsimacko(3903), felixcoto(3877), GreenTea(3867), Accoun(3846), rsergio(3818), Hazard(3765), rs5147(3753), joakim_mjardner(3734), dmj111(3720), shangas(3712) 2010-10-29T19:13:30 I added a distance table to the pastebin link 2010-10-29T19:13:34 anybody figure out why mega1 utterly destroys everyone yet lol 2010-10-29T19:13:51 he has a particular way of expansion that i can't quite figure out 2010-10-29T19:14:23 is he on tcp 2010-10-29T19:15:15 yep 2010-10-29T19:15:25 i noticed him there before official 2010-10-29T19:15:32 what's the bot name? 2010-10-29T19:15:37 one of my testBot incarnations got destroyed by him 2010-10-29T19:15:41 i was like who the hell of that 2010-10-29T19:15:43 is* 2010-10-29T19:15:50 bocsimacko 2010-10-29T19:15:54 oh 2010-10-29T19:15:55 him 2010-10-29T19:15:57 b.3x last i saw 2010-10-29T19:16:01 yeah, I think I know why 2010-10-29T19:16:17 he's got forecasting but that's not new 2010-10-29T19:16:23 i am trying to figure out his neutral selection 2010-10-29T19:16:26 McLeopold: the original is unchanged, you have to post the new one 2010-10-29T19:16:33 it's strange but not distance, not payback, not overall payback 2010-10-29T19:16:43 http://pastebin.com/ZZK3TkzF 2010-10-29T19:17:01 thanks 2010-10-29T19:17:02 *** Commandir has left #aichallenge 2010-10-29T19:17:27 overall payback is useless 2010-10-29T19:19:49 well there's 2 ways to look at individual planets imo 2010-10-29T19:19:58 how quickly it gets you ships, and how many ships you will get by x turn 2010-10-29T19:20:06 by payback i mean the former, only across the board 2010-10-29T19:20:21 so whatever order he gets the most return was my first thought 2010-10-29T19:20:33 *** davidjliu has joined #aichallenge 2010-10-29T19:20:52 overall payback isn't useless just has to be tempered against overall quantity 2010-10-29T19:21:31 okay, ships by x turn is useless 2010-10-29T19:21:39 except that's what wins games 2010-10-29T19:21:49 excepting annihilation ofc 2010-10-29T19:22:02 it isn't the end-all of scoring 2010-10-29T19:22:28 no but if you end up in a stalemate at the end and the other guy is pro and you lose by 5 it's because you didn't look at that 2010-10-29T19:23:31 you knapsack grows in size, therefore ships in x turns is useless 2010-10-29T19:25:01 i've given up on knapsack 2010-10-29T19:25:10 too concentrated on individual planets 2010-10-29T19:27:11 jmcarthur: have you finished your analysis? 2010-10-29T19:27:59 *** yasith has joined #aichallenge 2010-10-29T19:32:10 *** jesionaj` has quit IRC (Ping timeout: 255 seconds) 2010-10-29T19:32:30 *** jesionaj has joined #aichallenge 2010-10-29T19:35:24 *** yasith has quit IRC (Ping timeout: 252 seconds) 2010-10-29T19:41:18 *** jesionaj has quit IRC (Ping timeout: 252 seconds) 2010-10-29T19:45:43 *** bhasker_ has joined #aichallenge 2010-10-29T19:46:31 *** rangzen has quit IRC (Quit: Quitte) 2010-10-29T19:47:35 *** bhasker has quit IRC (Ping timeout: 240 seconds) 2010-10-29T19:48:46 *** yasith has joined #aichallenge 2010-10-29T19:49:09 *** mceier has quit IRC (Quit: leaving) 2010-10-29T19:55:36 *** McLeopold has quit IRC (Quit: Leaving) 2010-10-29T20:07:12 *** Mathnerd314 has quit IRC (Ping timeout: 245 seconds) 2010-10-29T20:11:57 *** demerzel_ has quit IRC (Quit: Page closed) 2010-10-29T20:12:41 Top 10 players: bocsimacko(3925), Accoun(3891), felixcoto(3876), GreenTea(3836), rsergio(3811), rs5147(3784), Hazard(3780), joakim_mjardner(3736), dmj111(3734), shangas(3722) 2010-10-29T20:37:53 *** boegel has joined #aichallenge 2010-10-29T20:38:41 *** Frontier has joined #aichallenge 2010-10-29T20:41:46 *** McLeopold has joined #aichallenge 2010-10-29T20:41:48 *** n9thbit has quit IRC (Quit: n9thbit) 2010-10-29T20:57:22 *** phreeza has quit IRC (Quit: Computer has gone to sleep) 2010-10-29T21:00:25 *** bhasker_ has quit IRC (Ping timeout: 272 seconds) 2010-10-29T21:02:06 *** demerzel_ has joined #aichallenge 2010-10-29T21:09:44 *** Todd200 has joined #aichallenge 2010-10-29T21:10:15 Hi people 2010-10-29T21:10:35 Yay, 3 servers up! 2010-10-29T21:12:41 Top 10 players: bocsimacko(3943), felixcoto(3933), Accoun(3914), GreenTea(3869), rsergio(3868), Hazard(3791), rs5147(3790), dmj111(3777), joakim_mjardner(3768), shangas(3747) 2010-10-29T21:12:54 Wow I stink at writing bots :/ 2010-10-29T21:13:52 I wrote a future predicting offense code, and it didn't work, anyone here? 2010-10-29T21:13:55 If you just started now you have a difficult time to get a decent score ^^ 2010-10-29T21:14:25 It doesn't matter, I'm building off of the starter pack anyhow 2010-10-29T21:14:47 * Todd200 is trying to learn from the experience 2010-10-29T21:15:32 But I can only minorly tweak the starting bot, or it will always crash 2010-10-29T21:15:33 That's a good way to learn. Test your bot again and again to find its weaknesses against other stronger bots. 2010-10-29T21:15:40 what lang 2010-10-29T21:15:47 Python 2010-10-29T21:15:52 oh the one i don't know haha 2010-10-29T21:16:04 Lol 2010-10-29T21:16:06 well i don't know lisp either but i'm counting major ones 2010-10-29T21:16:07 Todd200: learn to code then learn to write bots :p god damnit! 2010-10-29T21:16:17 i recommend rewriting the entire bot from scratch 2010-10-29T21:16:20 helped me a lot 2010-10-29T21:16:35 I doubt it would run 2010-10-29T21:16:36 stdio stuff is pretty easy shit in most languages but you get a feel for the data structure 2010-10-29T21:17:08 all you need for a running bot is something that reads 'go\n' and sends 'go\n' back 2010-10-29T21:17:12 go from there ;) 2010-10-29T21:17:53 * Todd200 doesn't understand the part that submits the move, but understands rest of bot 2010-10-29T21:18:28 Todd200: Your bot can crash when "FinishTurn" isn't called exactly once or you send more ships than you have on a planet or you send from a planet that isn't yours. Keep in mind that there are situations when either you or the enemy doesn't posess a single planet but the game is not over 2010-10-29T21:18:29 Hence why I can't rewrite it 2010-10-29T21:19:18 idk python standard input/output but speaking for C# and C++ both the console i/o is needlessly complex and pretty damn inefficient 2010-10-29T21:19:22 Irc glitch 2010-10-29T21:19:27 I'm blind 2010-10-29T21:19:55 can write a 2 line bot lol 2010-10-29T21:20:09 if (Console.Readline() == 'go\n') { Console.Writeline("go\n");} 2010-10-29T21:20:15 while(true) 2010-10-29T21:20:16 you win 2010-10-29T21:20:41 It wiill never lose to rage bot 2010-10-29T21:20:48 Amazing, you can also remove all whitespace from every c++ bot 2010-10-29T21:21:05 what 2010-10-29T21:21:10 Python is whitespase sensiitive 2010-10-29T21:21:18 oh wow that sucks 2010-10-29T21:21:22 i had my fill of that with php 2010-10-29T21:21:23 So if it mess up it crashes 2010-10-29T21:21:34 thus you cannot write a 2 MB program in a single line like in c++ 2010-10-29T21:21:43 c++ is somewhat whitespace sensitive though 2010-10-29T21:21:43 I love python besides that 2010-10-29T21:21:58 if you have multiple pointers and/or you use streams (i don't) 2010-10-29T21:22:04 i write C++ like oop C 2010-10-29T21:22:34 demerzel_: so you write imperative code using objects? 2010-10-29T21:22:47 Frontier: a single line program isn't really a plus 2010-10-29T21:23:04 i mean more that i don't use stdlib that much 2010-10-29T21:23:12 McLeopold: That statement wasn't opened to discussion. 2010-10-29T21:23:27 People write beautifiers so they can some of the crap out there 2010-10-29T21:23:38 i use stdio a lot more sprintf and c strings instead of std::string and <>holy<> weird < demerzel_: I like my 'log[info] << "This works like a charm!" << endl;' 2010-10-29T21:24:51 yea i use streams for file and console ;) that's about it tho 2010-10-29T21:25:48 Here's my offense booster code (that doesn't work) 2010-10-29T21:26:04 i am pretty proud of my c++ so far - managed to get an extracted state/supply/demand class working with oop 2010-10-29T21:26:10 use pastebin for code 2010-10-29T21:26:14 don't paste into the channel, please 2010-10-29T21:26:16 yea i'm hoping he's going to ahah 2010-10-29T21:26:19 http://pastebin.com/JBiAJXXw 2010-10-29T21:26:23 :) 2010-10-29T21:26:26 I know 2010-10-29T21:26:54 Why do people hate that? 2010-10-29T21:27:08 are you mixing up source/dest? 2010-10-29T21:27:15 iirc orders are source dest ships 2010-10-29T21:27:23 idk what the starter pack function is though 2010-10-29T21:27:41 I defined all the variables I needed.... 2010-10-29T21:27:46 Let's see 2010-10-29T21:27:57 i am referring to -- pw.issueorder(desti, source, num_ships) 2010-10-29T21:28:11 how does python work? every loop is a foreach and types are implicit? that's crazy 2010-10-29T21:28:18 or is there a mile long dec section somewhere 2010-10-29T21:28:37 The source was defined by the program earlier 2010-10-29T21:29:11 for some reason python is alien to me 2010-10-29T21:29:17 every other lang i can read and comprehend if not fully 2010-10-29T21:29:18 Source = p.planetid() 2010-10-29T21:29:38 I'm the opposite, I'm just bad at writing 2010-10-29T21:29:57 no i mean check the issueorder section 2010-10-29T21:30:10 you are giving it (dest, source, ships) i'm pretty sure it should be (source, dest, ships) 2010-10-29T21:30:18 if it's not that's bizarre because that's the order it prints it 2010-10-29T21:30:45 I'll check 2010-10-29T21:31:23 Nice, there's my first problem 2010-10-29T21:31:37 i'm not sure what you're trying to do there 2010-10-29T21:31:56 that's iterating your fleets in flight 2010-10-29T21:32:01 *** bhasker has joined #aichallenge 2010-10-29T21:32:09 you are adding defense per turns left and doing what with it? 2010-10-29T21:32:16 It looks at my fleets in flight, sees if they will kill the target, if not it sends extra 2010-10-29T21:32:32 ahhh ok 2010-10-29T21:32:36 that should work 2010-10-29T21:33:03 And it calculates how many ships the planet will have when my fleet arrives 2010-10-29T21:33:14 except you're not counting enemy fleets inbound 2010-10-29T21:33:18 you should add that 2010-10-29T21:33:38 I'll add defense after ofference starts working 2010-10-29T21:33:45 no i mean you're iterating turns 2010-10-29T21:33:55 check enemy fleets instead and just add growth * turns beforehand 2010-10-29T21:34:20 You're, right, I need to could their defense fleets 2010-10-29T21:34:32 i'd calc the growth in before i launched 2010-10-29T21:34:38 then you only need to check enemy fleets 2010-10-29T21:34:58 True, it would keep my loops smaller 2010-10-29T21:35:00 since their growth isn't going to change ;) 2010-10-29T21:35:26 As of right now it's sending 1/4 of it's strongest planet to the most cost effective planet 2010-10-29T21:36:04 So no pre-flight calculations 2010-10-29T21:36:15 *** seisatsu has joined #aichallenge 2010-10-29T21:36:36 http://ai-contest.com/profile.php?user_id=6765 2010-10-29T21:36:57 It's won 3 of it's 4 battles 2010-10-29T21:37:19 nice 2010-10-29T21:37:47 It still is firing blindly and doesn't care if it gets the planet it attacked or not 2010-10-29T21:37:55 sending a % of your source doesn't make sense for long though 2010-10-29T21:38:15 only thing that matters is whether you take something or not, you can just calc how much you need - (ships + (growth * distance)) + 1 2010-10-29T21:38:18 if you got it, send it 2010-10-29T21:38:38 excluding neutrals of course 2010-10-29T21:38:44 My bot will have a lot of if loops 2010-10-29T21:38:50 welcome lol 2010-10-29T21:39:26 i will tell you how i have done it since my first bot 2010-10-29T21:39:37 calc all pops for all planets for all turns from this turn on 2010-10-29T21:39:42 prior to starting my turn 2010-10-29T21:39:59 then i can simply just say do i have planet2[turn + distance] ships? if so send 2010-10-29T21:40:27 neutrals are just the same number x 200 2010-10-29T21:40:40 I was trying to add distance to my value of planet calcations, instant crash 2010-10-29T21:40:59 you should check that it's not the same index 2010-10-29T21:41:05 lavalamp_: is there go starter kit I can test with? 2010-10-29T21:41:13 ? 2010-10-29T21:41:15 i cache planet distances on turn 1 actually 2010-10-29T21:41:47 since they don't change and distances are incredibly useful 2010-10-29T21:41:59 i would say the most important number after 'ships' 2010-10-29T21:42:12 Then production 2010-10-29T21:42:36 ;) 2010-10-29T21:43:07 So I need to figure out distances 2010-10-29T21:43:48 production is 'ships' though 2010-10-29T21:43:55 just not ships you have now ;) 2010-10-29T21:44:03 True 2010-10-29T21:45:48 Couldn't you crash a futureseeing bot by using lots of fireworks? 2010-10-29T21:46:10 only if they are very poorly optimized hehe 2010-10-29T21:46:26 considering you only need to consider the enemy fleets once 2010-10-29T21:46:28 Mine would crash vs fireworks then 2010-10-29T21:47:13 why can you not click on names in tcp server argh 2010-10-29T21:47:15 The code I showed you looks at every fleet every turn 2010-10-29T21:47:21 Fireworks would destroy it 2010-10-29T21:47:32 nah i think you underestimate 2010-10-29T21:47:50 it's not efficient i agree but i doubt even 1000 iterations would bother a modern computer 2010-10-29T21:47:57 *** Cold-Phoenix has quit IRC () 2010-10-29T21:48:11 1 second limit 2010-10-29T21:49:07 I need a better offense system, instead of a strongest planet attack 2010-10-29T21:49:32 And I'm too noobish to code supply trains 2010-10-29T21:49:47 A thousand instructions per second (kIPS) is rarely used, as most current microprocessors can execute a least a billion instructions per second 2010-10-29T21:50:09 :/ 2010-10-29T21:50:19 now granted those are asm level but i doubt python is that horribly bloated 2010-10-29T21:50:24 supply trains are distance based 2010-10-29T21:50:34 I know 2010-10-29T21:50:53 read this thread 2010-10-29T21:50:54 http://www.ai-contest.com/forum/viewtopic.php?f=17&t=894 2010-10-29T21:51:01 it's pretty simple to make an easy one 2010-10-29T21:51:08 or rather the math behind it is simple 2010-10-29T21:51:37 I was thinking of an organized swarm instead of paths 2010-10-29T21:51:55 And an endgame code :) 2010-10-29T21:52:45 If I have 150% enemies ships and 125% enemies production rate, finish the job 2010-10-29T21:53:07 Or something of that idea 2010-10-29T21:54:49 I also need defense :/ 2010-10-29T21:55:23 And a 6 ship sniper would be cool 2010-10-29T22:00:27 *** Todd200 has quit IRC (Quit: thanks demerzel) 2010-10-29T22:08:09 aww look i helped 2010-10-29T22:12:41 Top 10 players: felixcoto(3933), bocsimacko(3920), Accoun(3887), GreenTea(3875), rsergio(3838), dmj111(3772), Hazard(3769), rs5147(3764), joakim_mjardner(3738), shangas(3731) 2010-10-29T22:26:28 *** Naktibalda_ has quit IRC (Remote host closed the connection) 2010-10-29T22:26:54 *** Palmik has quit IRC (Remote host closed the connection) 2010-10-29T22:49:22 *** roginn_ has joined #aichallenge 2010-10-29T22:53:46 *** TravisD has joined #aichallenge 2010-10-29T23:06:08 *** TravisD has quit IRC (Quit: TravisD) 2010-10-29T23:06:17 *** virdo has joined #aichallenge 2010-10-29T23:08:23 *** tapwater_ has joined #aichallenge 2010-10-29T23:10:15 *** Appleman1234 has quit IRC (Ping timeout: 272 seconds) 2010-10-29T23:10:57 *** tapwater has quit IRC (Ping timeout: 245 seconds) 2010-10-29T23:10:57 *** tapwater_ is now known as tapwater 2010-10-29T23:12:41 Top 10 players: felixcoto(3914), bocsimacko(3897), GreenTea(3858), Accoun(3841), rsergio(3810), rs5147(3765), dmj111(3752), Hazard(3730), joakim_mjardner(3726), shangas(3700) 2010-10-29T23:18:05 *** roginn_ has quit IRC (Remote host closed the connection) 2010-10-29T23:20:12 *** McLeopold has quit IRC (Ping timeout: 252 seconds) 2010-10-29T23:38:55 ahahaha holy shit i have an ungodly amount of state calcs etc and i'm benching at < 0.01s 2010-10-29T23:38:57 i love c++ 2010-10-29T23:39:11 err < 0.001 even 2010-10-29T23:57:02 *** yasith has quit IRC (Ping timeout: 265 seconds)