2010-09-20T00:00:00 deepblue: just curious, you don't have to answer, but do you have turn limit logic in your code? 2010-09-20T00:00:09 no 2010-09-20T00:00:18 mine just wants to win 2010-09-20T00:00:19 I saw a game where you had a hugh advantage but lost it after 200 2010-09-20T00:00:26 yeah 2010-09-20T00:00:29 that had other reasons ;) 2010-09-20T00:00:35 mine will have turn limit aware 2010-09-20T00:00:38 *will be 2010-09-20T00:00:49 i just don't like the idea 2010-09-20T00:00:54 to tailor it to the contest :/ 2010-09-20T00:00:57 me neither 2010-09-20T00:00:57 eventually all the top bots will be, i think 2010-09-20T00:01:08 me neither, but that's the way it will go 2010-09-20T00:01:15 Well, not number #1. :D 2010-09-20T00:01:17 IMO the turn limit is simply a part of the problem statement (and should therefore be nailed down) 2010-09-20T00:01:29 McLeopold: "eventually" 2010-09-20T00:01:47 yes, there is no way I can last 2 months. 2010-09-20T00:01:58 "have the most ships after turn 200" 2010-09-20T00:02:08 I just uploaded some fresh version to the official site 2010-09-20T00:02:12 let's see how it does :) 2010-09-20T00:02:26 I hope you go the the top again. I was dissapointed to see you fall. 2010-09-20T00:02:28 my friend had the idea 2010-09-20T00:02:41 of just doing very stupid things in the last turns when it's clear that you certainly win 2010-09-20T00:02:46 So, new stragety altogether? 2010-09-20T00:02:46 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T00:02:57 just tried some different variations 2010-09-20T00:03:00 and implementations 2010-09-20T00:03:16 *** danielvf has joined #aichallenge 2010-09-20T00:03:17 Dang, I was going to do that too. 2010-09-20T00:03:20 it's back to the old structure but with some optimizations until I get the branch working right :) 2010-09-20T00:03:40 what would be awesome 2010-09-20T00:03:44 I was going to try and lose with a bang. 2010-09-20T00:03:47 is to paint letters or something :) 2010-09-20T00:03:52 lol 2010-09-20T00:03:53 Yes 2010-09-20T00:03:55 "gg" 2010-09-20T00:04:16 like a signature 'good bye' 2010-09-20T00:04:19 hehe 2010-09-20T00:04:37 I'm going to write one now and put it on the tcp server. 2010-09-20T00:04:48 http://ai-contest.com/visualizer.php?game_id=4595271 2010-09-20T00:04:51 yay 2010-09-20T00:04:54 first game a win 2010-09-20T00:04:59 the server is quick 2010-09-20T00:05:37 Yea, you get your first game right away. 2010-09-20T00:05:40 awesome, you beat someone ranked ~1000 2010-09-20T00:05:41 :P 2010-09-20T00:05:51 amazing isn't it? 2010-09-20T00:05:52 ;) 2010-09-20T00:06:10 but he/she is from waterloo! 2010-09-20T00:06:16 wtf, your rank is 85 after that one game! 2010-09-20T00:06:19 jmcarthur: any chance of convincing the powers that be that turn limit is not part of the problem descriptions? 2010-09-20T00:06:27 no that's my old rank 2010-09-20T00:06:34 oh 2010-09-20T00:06:51 990? 2010-09-20T00:06:52 i was going to say, it took my current bot like 6 wins to get that high 2010-09-20T00:07:02 ah, now it's 990 2010-09-20T00:07:19 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T00:07:21 yaya 2010-09-20T00:07:31 that will be a long climb again 2010-09-20T00:07:38 poor guy, he got demolished by someone ranked 990... 2010-09-20T00:07:46 hehe 2010-09-20T00:07:53 and he's off to a bad start 2010-09-20T00:07:56 first two games he lost 2010-09-20T00:07:58 everytime you win, his rank will now rise. 2010-09-20T00:08:17 lucky him 2010-09-20T00:08:19 http://ai-contest.com/visualizer.php?game_id=4595207 second fight of my boty 2010-09-20T00:08:24 this ranking system is bad for beginners in that way 2010-09-20T00:08:56 your's is going straight forward 2010-09-20T00:09:09 mine is always like 2010-09-20T00:09:24 'look at the landscape! and there what a nice planet let's check it out' 2010-09-20T00:10:22 I get a connection refused from the tcp server 2010-09-20T00:10:26 *** danielvf has joined #aichallenge 2010-09-20T00:11:24 Same here 2010-09-20T00:12:02 *** sannysanoff has quit IRC (Ping timeout: 276 seconds) 2010-09-20T00:12:05 I was getting that too, now it just won't play. 2010-09-20T00:12:13 *** Zannick has quit IRC (Quit: leaving) 2010-09-20T00:12:22 *** clever_name has joined #aichallenge 2010-09-20T00:14:14 *** danielvf_ has joined #aichallenge 2010-09-20T00:14:14 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T00:14:14 *** danielvf_ is now known as danielvf 2010-09-20T00:15:21 dhartmei_: Do you mind if I put your server addr as the default in my gui tcp lient? 2010-09-20T00:15:24 er, client 2010-09-20T00:15:34 def ranked 1677 right now after 2 wins. hm... 2010-09-20T00:15:57 greghaynes: he said he was moving servers 2010-09-20T00:16:10 Yea, I got the new addr 2010-09-20T00:18:42 bah, act first, ask later ;) 2010-09-20T00:19:12 greghaynes: use @later tell dhartmei_ 2010-09-20T00:19:21 Hes here, hell get it 2010-09-20T00:19:39 greghaynes: backlog for this channel is huge 2010-09-20T00:19:40 *** JCS^ has quit IRC (Quit: \x00) 2010-09-20T00:19:51 Thats why you use /last ;) 2010-09-20T00:20:04 darn, i don't have that 2010-09-20T00:21:25 hrm, the new address doesnt seem to work 2010-09-20T00:22:12 Are others able to play on benzedrine at the moment? 2010-09-20T00:22:34 i appear to be playing on there now 2010-09-20T00:22:34 greghaynes: It just started working here. 2010-09-20T00:22:34 should we be connecting to the 72.44... address? 2010-09-20T00:22:47 72.44.46.68, yes 2010-09-20T00:22:58 not working for me... 2010-09-20T00:23:06 hrm, is there a new protocol or something? 2010-09-20T00:23:16 i just pasted 72.44.46.68 into vim 2010-09-20T00:23:23 works again for me 2010-09-20T00:23:36 greghaynes: do you have the version that supports -p password? 2010-09-20T00:23:52 This is for my own client, but I support that... 2010-09-20T00:25:01 gah, dumb mistake 2010-09-20T00:25:02 Wrong port 2010-09-20T00:26:15 yeah, me to, I just needed to RTFM 2010-09-20T00:28:05 map111.txt? 2010-09-20T00:28:22 yeah, 100 extra maps were added 2010-09-20T00:28:34 what's the new site 2010-09-20T00:28:36 i wanna see it too 2010-09-20T00:29:23 amstan: see what? 2010-09-20T00:29:40 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T00:29:54 new tcp server 2010-09-20T00:29:59 lol, who has the ragebot running on the tcp? :) 2010-09-20T00:30:07 where do I get the new maps? 2010-09-20T00:30:14 amstan: links from http://www.benzedrine.cx/planetwars/ go to the correct place 2010-09-20T00:30:24 McLeopold: http://72.44.46.68/getmaps 2010-09-20T00:30:29 aww 2010-09-20T00:30:37 still old interface 2010-09-20T00:30:58 html 3 site 2010-09-20T00:31:19 sigh: is there a zip file or something? 2010-09-20T00:31:24 dunno 2010-09-20T00:31:45 I'd just use a wget loop 2010-09-20T00:32:07 Why the new map? 2010-09-20T00:32:11 maps 2010-09-20T00:32:16 I don't know how to write a loop. 2010-09-20T00:32:20 :) 2010-09-20T00:32:22 ha ha 2010-09-20T00:32:29 :P 2010-09-20T00:32:29 just kidding 2010-09-20T00:33:19 genericbob: for variety? 2010-09-20T00:33:52 are there any mirror image maps? 2010-09-20T00:34:50 I created some, but I don't think they are on any servers. 2010-09-20T00:34:54 I'm guessing these were generated with the official script, in which case, no 2010-09-20T00:35:22 but yeah, genericbob's been talking about his uber-generic map generating scripts :D 2010-09-20T00:36:00 It would be fun to have non-semetrical maps. Everytime a side losses, modify the map to give that side 1 more starting fleet. 2010-09-20T00:36:10 mirrored maps are for pu**ies ;) 2010-09-20T00:36:15 Until you find a maps equalibrium. 2010-09-20T00:36:34 I suggested just play those games twice but everyone hated on me - shhhh 2010-09-20T00:36:39 yes :) 2010-09-20T00:36:48 I like playing twice 2010-09-20T00:36:54 I think you'd have a tie everytime... 2010-09-20T00:37:13 One win for each side, every time. 2010-09-20T00:37:16 only if they are really off 2010-09-20T00:37:20 true. 2010-09-20T00:37:29 and all the bots equally strong 2010-09-20T00:37:39 but then that also allows more strategy to 2010-09-20T00:37:48 the really strong bots then can make their points on the symmetric ones 2010-09-20T00:38:07 yes equaling less games 2010-09-20T00:38:38 Hmm. New entry must beat rage bot on a map in there favor... other wise their entry is rejected. 2010-09-20T00:38:47 And everyone else gets more games! 2010-09-20T00:39:01 [sic]their 2010-09-20T00:39:08 everybot has to win against ragebot on 30 2010-09-20T00:39:18 30 is evil 2010-09-20T00:39:22 hehe 2010-09-20T00:39:39 The first time I lost on 30 I crapped my pants. 2010-09-20T00:39:44 :) 2010-09-20T00:39:48 ;) 2010-09-20T00:40:02 I was like wtf when a1k0n posted on the forums that he was 100/100 against ragebot 2010-09-20T00:40:03 it's part of my standard testing routine 2010-09-20T00:40:11 I thought of writing, if round < 1, do nothing, just for that. 2010-09-20T00:40:17 hehe 2010-09-20T00:40:17 but then he updated later with something lame like 87/100 2010-09-20T00:40:21 *** HakanD has quit IRC (Ping timeout: 245 seconds) 2010-09-20T00:40:42 letting my bot run on all maps against all bots 2010-09-20T00:40:45 takes forever 2010-09-20T00:40:49 he had me seriously scared that everyone had uber bots 2010-09-20T00:40:51 97/100 for both for me 2010-09-20T00:40:54 since I use 800ms 2010-09-20T00:40:55 :/ 2010-09-20T00:41:10 deepblue: I run the games concurrently across my 4 cores 2010-09-20T00:41:11 Doesn't everyone beat all starter pack bots? 2010-09-20T00:41:14 100/100 for all of the example bots. 2010-09-20T00:41:24 McLeopold: I don't have 100% against rage 2010-09-20T00:41:27 I beat all :) 2010-09-20T00:41:37 which maps do you lose? 2010-09-20T00:41:58 my first version had a 2010-09-20T00:42:01 not sure... I'm not at home so I can't look it up 2010-09-20T00:42:05 if round = 0 return 2010-09-20T00:42:06 :D 2010-09-20T00:42:21 that'll net you a draw 2010-09-20T00:42:27 no 2010-09-20T00:42:39 as soon as he saw what was coming in round 1 2010-09-20T00:43:10 my current version manages just fine without that 2010-09-20T00:43:19 he survived it correctly 2010-09-20T00:44:30 map 33 against dualbot 2010-09-20T00:44:50 just running right now 2010-09-20T00:45:04 does anyone have a simple map viewer? 2010-09-20T00:45:35 Be more specific 2010-09-20T00:45:39 i just let my bot play against dualbot on 33 now just to see the map :D 2010-09-20T00:45:41 map 73 was a lost 2010-09-20T00:45:42 McLeopold: i am of the opinion that there *should* be a (concretely specified) turn limit 2010-09-20T00:45:47 um, view a map without playing a game 2010-09-20T00:46:00 *** nullkuhl_ is now known as nullkuhl 2010-09-20T00:46:06 map 84 2010-09-20T00:46:23 jmcarthur: are you sure you wouldn't rather have end game conditions similar to chess? 2010-09-20T00:47:19 *** antimatroid1 has joined #aichallenge 2010-09-20T00:47:27 jmcarthur: must be forward progress, otherwise a draw 2010-09-20T00:47:43 McLeopold: i don't think that would be sufficient to prevent infinite games 2010-09-20T00:47:56 I chagned something - reserveed ships - ragebot is beating my current bot all the time 2010-09-20T00:47:57 McLeopold: unless "forward progress" is prohibitively strict 2010-09-20T00:48:37 jmcarthur: your right, it would probably have to be. My only thought was 200 turns after the last neutral take over. 2010-09-20T00:48:43 *** antimatroid has quit IRC (Ping timeout: 240 seconds) 2010-09-20T00:48:54 97/100 dual 65 turns, 88/100 rage 56 turns (need to go fix that) 2010-09-20T00:49:11 jmcarthur: that would actually be pretty easy to code for, and test for on the server 2010-09-20T00:49:22 McLeopold: what if the last neutral is never taken over? 2010-09-20T00:49:33 200 turns since last planet taken 2010-09-20T00:49:44 jmcarthur: I mean, if a neutral hasn't been taken for 200 turns, end the game 2010-09-20T00:49:51 ah 2010-09-20T00:50:05 max game time would still be very large (albeit rare) 2010-09-20T00:50:11 200 turns isnt enough in a close game 2010-09-20T00:50:20 *** danielvf has joined #aichallenge 2010-09-20T00:50:27 199*(num_planets - 2) + 1 2010-09-20T00:50:31 yeah, maybe at least 200, 10 extra after the last neutral? 2010-09-20T00:50:35 if I have 2 more growth then you and you have 500 more ships then me I will never win 2010-09-20T00:50:53 Unless my bot goes catatonic? 2010-09-20T00:51:09 Were not writing perfect bots here... 2010-09-20T00:51:10 McLeopold: i don't think that would actually change the outcome of any of the games you have used as examples, would it? 2010-09-20T00:51:32 its not about perfect bots, its about if Im gaining ground and only lose because of a turn limit - kinda sucks 2010-09-20T00:51:37 jmcarthur: it's hard to tell, I can't query the raw data 2010-09-20T00:51:54 Cyndre: agreed 2010-09-20T00:52:02 *** Zannick has joined #aichallenge 2010-09-20T00:52:08 http://72.44.46.68/canvas?game_id=51829 2010-09-20T00:52:18 i once considered x turns after the last "cross-over", where a cross-over is a point where the upper hand switches sides, but even that isn't sufficient (when bots simply switch back and forth repeatedly) 2010-09-20T00:52:37 *** syntaxglitch has quit IRC (Read error: Operation timed out) 2010-09-20T00:52:46 http://72.44.46.68/canvas?game_id=52004 2010-09-20T00:52:49 thats a good one 2010-09-20T00:52:55 honestly if you guys "only lose because of a turn limit" then it's because you weren't playing by the same rules as everybody else: "have the most ships after 200 turns" 2010-09-20T00:52:56 I think x turns after the last neutral takeover would be best, if anything. Eventually you would run out and i'd have to end. 2010-09-20T00:53:24 i'm fine with the objective not being "completely kill the other player" 2010-09-20T00:54:42 not sure why my bot was keeping that big bunch of planets around 2010-09-20T00:54:58 jmcarthur: what about a large margin for draws? what's your option? 2010-09-20T00:55:19 http://72.44.46.68/canvas?game_id=51567 or look at this one - we were deadlocked and then bang he made a mistake 2010-09-20T00:55:31 well, if we say the objective is as i've stated it then it makes sense for draws to only be when ships are exactly the same 2010-09-20T00:55:47 what about growth rate? 2010-09-20T00:55:52 and if it's a win by only 1 ship, well, that bot planned better 2010-09-20T00:56:16 growthRate, planetsControlled, and Ships should be compared 2010-09-20T00:56:16 Top 10 players: McLeopold(3610), kcm1700(3451), EBraun(3435), paulus(3386), adumlah(3353), golfin(3326), sighbot(3316), kevlar(3277), wagstaff(3247), Titannet1(3228) 2010-09-20T00:56:32 factoring in growth rate just muddies the rules. doesn't necessarily make the game more interesting 2010-09-20T00:56:35 IMO 2010-09-20T00:56:54 playing for a draw might make it more interesting 2010-09-20T00:57:04 you can still play for a draw 2010-09-20T00:57:14 and if it's difficult it means you are far enough behind to lose anyway 2010-09-20T00:57:20 but it's like trying to bring down the death star 2010-09-20T00:57:33 right. you lost :P 2010-09-20T00:57:34 but the rebels still win 2010-09-20T00:58:05 yea but if I have higher growth rate and nothing changes I win 2010-09-20T00:58:12 *** AlliedEnvy has quit IRC (Ping timeout: 276 seconds) 2010-09-20T00:58:13 so, at this point is it just trying to find the magic turn limit number? 2010-09-20T00:58:39 but if its a close game - a few growthRate in difference, 200 turns is only 400 ships 2010-09-20T00:58:45 at 2 growthRate difference 2010-09-20T00:58:54 It would be nice to see some stats on average game length per player 2010-09-20T00:58:55 at which point each side could have 2000 ships 2010-09-20T00:59:00 i think changing the endgame conditions is just arbitrary and a matter of opinion, not really a game changer (except for making turn limit variable or nonexistent, in which case many algorithms become invalid) 2010-09-20T00:59:06 http://ai-contest.com/visualizer.php?game_id=4581455 2010-09-20T00:59:09 theres a good one 2010-09-20T00:59:18 s/endgame/win-lose/ 2010-09-20T00:59:48 that game is clearly not over :) 2010-09-20T00:59:56 my bot, for example, relies critically on the existence of a concrete turn count 2010-09-20T01:00:32 that game is over because the rules (should) say it's over 2010-09-20T01:00:40 and the winner won because he had the upper hand 2010-09-20T01:00:45 at the right time 2010-09-20T01:00:50 nope, that game wasn't over at all 2010-09-20T01:00:58 or have a player global turn count - you get 200 turns per game and if you dont use them all they goto a pool that you can borrow from 2010-09-20T01:01:07 except for the rules, yes 2010-09-20T01:01:16 lmao 2010-09-20T01:01:33 http://ai-contest.com/visualizer.php?game_id=4537465 or that one 2010-09-20T01:01:37 I'd like to send ships from my losing games into the start of my other games... 2010-09-20T01:01:47 *** AlliedEnvy has joined #aichallenge 2010-09-20T01:01:57 basketball games are won by single points sometimes. they don't get overtime just for being close but not drawing 2010-09-20T01:02:10 good point 2010-09-20T01:02:13 but 200 turns isnt enough for a difference to be seen 2010-09-20T01:02:24 not so much in baseball 2010-09-20T01:02:37 nor football (american) 2010-09-20T01:02:39 Cyndre: we haven't made 200 the final number yet 2010-09-20T01:02:51 i'm all for making the number bigger 2010-09-20T01:02:54 nor finals in the world cup 2010-09-20T01:02:57 i simply think it should be fixed 2010-09-20T01:03:04 for a good example go look at McLeopolds losing games, everyone is at the limit 2010-09-20T01:03:17 *** Zannick has quit IRC (Read error: Connection reset by peer) 2010-09-20T01:03:20 McLeopold: sure, *draws* are drug out like that, but simply close games aren't 2010-09-20T01:03:24 jmcarthur: could you tell from the games played so far what an upper bound might be? 2010-09-20T01:03:36 McLeopold: not yet 2010-09-20T01:03:37 jmcarthur: wait, you can't because you cut them off... 2010-09-20T01:04:00 what you should do is have a way to pit two bots on map y or an admin can replay a game for longer turns 2010-09-20T01:04:06 it would be nice to have started longer and cut it as time goes by rather than the other way, i agree 2010-09-20T01:04:38 some bots use random, so you can't replay 2010-09-20T01:04:43 4/5 of McLeopolds losses were at 200 turns 2010-09-20T01:05:27 i just don't think we need some stronger sense of "closure" in a close game that finished due to a turn limit if we just accept that the objective is to have the most ships at a particular turn 2010-09-20T01:05:50 but my argument is if you dont get a good start, but catch them in a mistake there sno point 2010-09-20T01:05:51 once I accept that, I will agree. I'm not there yet. :) 2010-09-20T01:06:03 because you cant win in that small ammount of turns 2010-09-20T01:06:28 say turn 50 they are ahead by 300 ships and on turn 100 you catch them in a mistake and get 1 more growthRate per turn - pointless 2010-09-20T01:06:30 Cyndre: i'm not arguing about the number at all 2010-09-20T01:06:41 Cyndre: i want a larger turn limit too 2010-09-20T01:06:50 Cyndre: as long as there *is* one 2010-09-20T01:06:51 Cyndre: it might come down to having to take more risk the choose the slow but sure way. 2010-09-20T01:06:55 jmcarthur: but were talking orders of magnitude higher - which the server cant handle 2010-09-20T01:07:01 *** DjinnKahn has quit IRC (Ping timeout: 252 seconds) 2010-09-20T01:07:07 Cyndre: we are getting new servers soon i think 2010-09-20T01:07:10 to try to balance the load 2010-09-20T01:07:15 cool :) 2010-09-20T01:07:30 there are at least very serious attempts to get that working 2010-09-20T01:07:37 *** Zannick has joined #aichallenge 2010-09-20T01:08:00 jmcarthur: thanks for discussing, I appreciate that 2010-09-20T01:08:02 but not even sure why you need a turn limit 2010-09-20T01:08:11 i'm not sure about orders of magnitude though. do you really think games need to be 2000 or 20000 turns long? 2010-09-20T01:08:30 Cyndre: there are dumb bots that will dance forever. We need something. 2010-09-20T01:08:37 Cyndre: because we need to prevent infinite games and any other rules would be more complicated and controversial even than this 2010-09-20T01:08:40 Mcleopold: yea :( 2010-09-20T01:08:55 give me 2k turns max lol 2010-09-20T01:09:18 2000, yes, 20000, probably no. It's really a function of the map sizes. 2010-09-20T01:09:25 yea 2010-09-20T01:09:31 which we should probably limit to 2010-09-20T01:09:32 o 2010-09-20T01:10:07 or make a scaling one based on number of time syou hit the max it scales it back on youi 2010-09-20T01:10:15 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T01:10:26 well, a map that can't be traversed during the entire game would be silly 2010-09-20T01:10:32 *** danielvf has joined #aichallenge 2010-09-20T01:10:34 so yes, there needs to be some limit 2010-09-20T01:10:35 but awesome 2010-09-20T01:10:38 yeah that just becomes arbitrary (plus, what about your opponents? how are they affected by your personal turn limit?) 2010-09-20T01:10:56 the game goes to the longest players turn limit 2010-09-20T01:10:58 jmcarthur: what do you mean? 2010-09-20T01:11:01 personal turn limit? 2010-09-20T01:11:04 sigh: i was responding to Cyndre 2010-09-20T01:11:14 ah, ok 2010-09-20T01:11:25 no, constant set rules is best 2010-09-20T01:11:39 I agree on that one 2010-09-20T01:11:50 okay, i'm out, cyl 2010-09-20T01:11:51 *** naglijo_ has quit IRC (Ping timeout: 276 seconds) 2010-09-20T01:11:55 cya 2010-09-20T01:11:56 *** McLeopold has quit IRC (Quit: Leaving) 2010-09-20T01:11:57 see you 2010-09-20T01:12:19 *** alexshendi has left #aichallenge ("Verlassend") 2010-09-20T01:12:26 it would be nice if the turns remaning were included in the game state format. Janzert submitted a patch to include that as a comment, iirc 2010-09-20T01:12:29 *remaining 2010-09-20T01:13:17 * Janzert nods, see issue 74 :) 2010-09-20T01:14:45 *** danielvf_ has joined #aichallenge 2010-09-20T01:14:54 *** sandeep_ has joined #aichallenge 2010-09-20T01:14:57 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T01:14:58 *** danielvf_ is now known as danielvf 2010-09-20T01:15:04 your tcp rocks - ty 2010-09-20T01:15:48 hi 2010-09-20T01:15:57 is any of the organizers here 2010-09-20T01:15:59 ? 2010-09-20T01:17:20 *** tapwater has quit IRC (Quit: tapwater) 2010-09-20T01:17:31 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T01:17:45 dang i'm stuck with a difficult decision. should i use a really nice library that only works with GHC HEAD in the hopes that we are allowed to use binaries eventually, or should i avoid such silliness? 2010-09-20T01:17:49 *** danielvf has joined #aichallenge 2010-09-20T01:18:00 avoid such silliness 2010-09-20T01:18:01 it would fundamentally affect the structure of some core algorithms of my bot 2010-09-20T01:18:11 but it would be so *fast* 2010-09-20T01:18:13 which is important 2010-09-20T01:18:20 the real question is do you want to seriously compete or compete for fun? 2010-09-20T01:18:27 both of course! 2010-09-20T01:18:50 no no, competing for fun can be done on tcp server, but you cant seriously compete on the final server 2010-09-20T01:18:58 if your bot doesnt compile 2010-09-20T01:19:33 which is pretty lame :( 2010-09-20T01:19:34 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T01:20:01 *** hornairs has quit IRC (Remote host closed the connection) 2010-09-20T01:20:06 *** danielvf has joined #aichallenge 2010-09-20T01:20:18 another thing about this is a get parallelism basically for free if i want it, which would give me an edge on the tcp server. i probably would intentionally *not* use that though 2010-09-20T01:20:31 *is i get 2010-09-20T01:21:34 jmcarthur: haha, just have an uber powerful bot to demoralse everyone on tcp 2010-09-20T01:21:39 heh 2010-09-20T01:21:45 that would be soooo lame :P 2010-09-20T01:21:54 run it under multiple different names so it looks like a bunch of strong people 2010-09-20T01:22:00 i could run it on a 16 core monster or something 2010-09-20T01:22:11 well, if only i had one 2010-09-20T01:22:16 :P 2010-09-20T01:22:34 I have a monster quad core with 8 gigs of ram running windows..... 2010-09-20T01:22:52 ew :P 2010-09-20T01:22:58 oh wait, thats at work, core 2 duo at home 2010-09-20T01:23:19 I hav e a E7500, 2.93GHz, 8.00 GB ram, 64 bit windows 7 key sike 2010-09-20T01:24:43 i should write a client with MPI support for my work cluster 2010-09-20T01:25:21 one second on 88 cores should be enough for a start 2010-09-20T01:26:30 *** Olathe has quit IRC (Ping timeout: 264 seconds) 2010-09-20T01:28:07 *** blbrown_win3 has quit IRC (Read error: Connection reset by peer) 2010-09-20T01:28:16 *** danielvf has quit IRC (Ping timeout: 245 seconds) 2010-09-20T01:32:19 *** Satya has quit IRC (Ping timeout: 252 seconds) 2010-09-20T01:34:51 *** danielvf has joined #aichallenge 2010-09-20T01:37:04 wb 2010-09-20T01:37:04 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T01:38:47 *** Eiii has quit IRC (Ping timeout: 240 seconds) 2010-09-20T01:39:16 *** danielvf has joined #aichallenge 2010-09-20T01:42:13 *** clever_name has quit IRC (Ping timeout: 252 seconds) 2010-09-20T01:43:00 *** danielvf_ has joined #aichallenge 2010-09-20T01:43:00 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T01:43:00 *** danielvf_ is now known as danielvf 2010-09-20T01:43:32 *** sandeep_ has quit IRC (Quit: Page closed) 2010-09-20T01:44:47 *** zero has joined #aichallenge 2010-09-20T01:44:56 . 2010-09-20T01:45:15 *** zero is now known as Guest47727 2010-09-20T01:46:10 *** AlliedEnvy has quit IRC (Ping timeout: 272 seconds) 2010-09-20T01:48:58 *** AlliedEnvy has joined #aichallenge 2010-09-20T01:51:16 *** deepblue has quit IRC (Quit: Page closed) 2010-09-20T01:53:53 *** danielvf_ has joined #aichallenge 2010-09-20T01:53:53 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T01:53:53 *** danielvf_ is now known as danielvf 2010-09-20T01:54:57 *** Guest47727 has quit IRC (Quit: Leaving) 2010-09-20T01:56:17 Top 10 players: McLeopold(3606), kcm1700(3446), EBraun(3432), sighbot(3341), adumlah(3338), paulus(3326), golfin(3310), kevlar(3297), Titannet1(3241), wagstaff(3239) 2010-09-20T01:56:45 *** MetroStyles has joined #aichallenge 2010-09-20T01:56:54 *** lemonedo has joined #aichallenge 2010-09-20T02:18:21 *** drewrobb has quit IRC (Read error: Connection reset by peer) 2010-09-20T02:28:03 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T02:28:07 *** lemonedo_ has joined #aichallenge 2010-09-20T02:28:44 *** lemonedo has quit IRC (Ping timeout: 240 seconds) 2010-09-20T02:28:52 *** danielvf has joined #aichallenge 2010-09-20T02:29:54 yarrr 2010-09-20T02:33:14 *** TapewormGalactos has joined #aichallenge 2010-09-20T02:34:00 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T02:34:52 *** danielvf has joined #aichallenge 2010-09-20T02:46:29 *** evandijk70 has joined #aichallenge 2010-09-20T02:47:08 *** Theobon has joined #aichallenge 2010-09-20T02:47:43 *** evandijk70 has quit IRC (Client Quit) 2010-09-20T02:48:19 *** lemonedo_ has quit IRC (Remote host closed the connection) 2010-09-20T02:49:05 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T02:49:16 *** danielvf has joined #aichallenge 2010-09-20T02:49:54 *** mega1 has joined #aichallenge 2010-09-20T02:53:52 *** lemonedo has joined #aichallenge 2010-09-20T02:56:17 Top 10 players: McLeopold(3606), kcm1700(3462), EBraun(3452), adumlah(3343), sighbot(3341), golfin(3321), paulus(3300), kevlar(3293), wagstaff(3246), Titannet1(3231) 2010-09-20T02:58:33 *** victor_ has joined #aichallenge 2010-09-20T02:58:49 how play galcon???? 2010-09-20T02:58:58 Well, thank you 2010-09-20T03:02:27 ???? 2010-09-20T03:03:09 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T03:03:22 *** danielvf has joined #aichallenge 2010-09-20T03:03:45 *** lemonedo has quit IRC (Remote host closed the connection) 2010-09-20T03:05:05 did tcp games on the site stop loading for anyone else? 2010-09-20T03:05:57 antimatroid1, yep 2010-09-20T03:11:35 *** dhartmei has joined #aichallenge 2010-09-20T03:12:52 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T03:13:32 *** danielvf has joined #aichallenge 2010-09-20T03:18:06 *** fisxoj has quit IRC (Ping timeout: 276 seconds) 2010-09-20T03:19:32 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T03:19:52 *** danielvf has joined #aichallenge 2010-09-20T03:21:32 *** victor_ has quit IRC (Quit: Page closed) 2010-09-20T03:25:47 *** danielvf has quit IRC (Ping timeout: 245 seconds) 2010-09-20T03:26:27 *** danielvf has joined #aichallenge 2010-09-20T03:27:35 *** Utkarsh has quit IRC (Ping timeout: 240 seconds) 2010-09-20T03:27:35 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T03:27:47 *** danielvf has joined #aichallenge 2010-09-20T03:33:42 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T03:33:45 *** danielvf_ has joined #aichallenge 2010-09-20T03:36:50 *** rwa has joined #aichallenge 2010-09-20T03:38:27 *** sigh has quit IRC (Remote host closed the connection) 2010-09-20T03:40:47 *** danielvf has joined #aichallenge 2010-09-20T03:40:53 *** danielvf_ has quit IRC (Read error: Connection reset by peer) 2010-09-20T03:41:39 *** delt0r___ has quit IRC (Read error: Operation timed out) 2010-09-20T03:42:14 *** hellman has joined #aichallenge 2010-09-20T03:42:17 *** delt0r___ has joined #aichallenge 2010-09-20T03:50:02 *** danielvf_ has joined #aichallenge 2010-09-20T03:50:02 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T03:50:03 *** Naith has quit IRC (Quit: Naith) 2010-09-20T03:50:04 *** danielvf_ is now known as danielvf 2010-09-20T03:54:04 *** Theobon has quit IRC (Ping timeout: 265 seconds) 2010-09-20T03:56:18 Top 10 players: McLeopold(3623), kcm1700(3483), EBraun(3474), adumlah(3379), sighbot(3364), kevlar(3337), golfin(3329), paulus(3292), wagstaff(3268), Titannet1(3240) 2010-09-20T04:02:43 *** wh1teside has joined #aichallenge 2010-09-20T04:16:24 *** z-247 has joined #aichallenge 2010-09-20T04:19:14 rwa: that's a mighty big elo score you've got there 2010-09-20T04:19:24 so far so good 2010-09-20T04:19:33 not if i can help it 2010-09-20T04:19:54 damn aha 2010-09-20T04:19:58 :) 2010-09-20T04:20:40 i keep making stupid mistakes and ruining my score 2010-09-20T04:20:45 but i don't think that was one of those times 2010-09-20T04:21:24 I think my bot likes that board 2010-09-20T04:21:34 ha, it was, but doesn't mean it's not going to be better than me 2010-09-20T04:21:43 i f'd up me redistribution completely 2010-09-20T04:22:23 *** TapewormGalactos has quit IRC (Quit: Ex-Chat) 2010-09-20T04:22:32 my bot is doing weird stuff. I think it's still buggy. 2010-09-20T04:22:38 either that or it's smarter than I am. 2010-09-20T04:22:43 probably the former. ;P 2010-09-20T04:22:49 *** mag has joined #aichallenge 2010-09-20T04:23:05 better go to bed before it starts losing 2010-09-20T04:23:08 night 2010-09-20T04:23:16 *** mag is now known as Guest81441 2010-09-20T04:24:18 night 2010-09-20T04:24:24 mine is being buggy now 2010-09-20T04:24:26 working it out 2010-09-20T04:24:32 leave yours running for a bit at least :P 2010-09-20T04:24:35 i want to beat it 2010-09-20T04:24:52 leaving it up overnight... 2010-09-20T04:26:41 cool 2010-09-20T04:28:28 rwa: got you 2010-09-20T04:28:55 nice 2010-09-20T04:30:14 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T04:30:31 is there any discussion regarding movement modelling for game tree generation? 2010-09-20T04:30:48 *** danielvf has joined #aichallenge 2010-09-20T04:31:23 good luck making a game tree that does anything useful 2010-09-20T04:33:25 antimatroid1: I think our bots are thinking along similar lines. But your implementation looks better so far. 2010-09-20T04:35:23 i'm still playing around with parameters, then i need to clean everything up 2010-09-20T04:35:49 it's still largely heuristic based, i'm finding it hard to work out how best to use anything beyond that without refining my heuristics better 2010-09-20T04:37:53 It seems doubtful that you'll find much rigorous "solving" in this game. 2010-09-20T04:38:11 i don't doubt the winner will have some nice little trick (or quite a few) though 2010-09-20T04:38:17 *** sigh has joined #aichallenge 2010-09-20T04:38:30 i have a few ideas, but they're gonna take a bit of effort 2010-09-20T04:38:38 put your bot up again sigh, we have new challengers 2010-09-20T04:38:57 the official site sucks at ranking bots this time 2010-09-20T04:40:47 *** nullkuhl has quit IRC (Ping timeout: 240 seconds) 2010-09-20T04:44:11 rwa: ok, they are started again 2010-09-20T04:44:17 I don't think they will do as well now 2010-09-20T04:44:54 sigh: my new bot seems to be much better 2010-09-20T04:45:03 antimatroid1: cool :) 2010-09-20T04:45:03 i look forward to pantsing you :) 2010-09-20T04:45:07 heh 2010-09-20T04:46:30 here we go again anti 2010-09-20T04:47:18 damn you aha 2010-09-20T04:47:22 booya 2010-09-20T04:47:23 long game 2010-09-20T04:47:30 and again 2010-09-20T04:47:47 fairly easy win in the end 2010-09-20T04:48:09 antimatroid1: ah, you've got redistribution code as well now 2010-09-20T04:48:27 i do, although it's got an obvious new stupid error :P, but a loss is a loss 2010-09-20T04:48:35 my bot's behavior looked buggy in that game 2010-09-20T04:49:45 sorry, needs fixing 2010-09-20T04:50:20 doh! 2010-09-20T04:50:23 horrible bug... 2010-09-20T04:51:16 will restart in a moment 2010-09-20T04:51:18 *** kozlovsky has joined #aichallenge 2010-09-20T04:51:23 (watch, probably get worse) 2010-09-20T04:51:26 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T04:52:02 *** danielvf has joined #aichallenge 2010-09-20T04:52:56 alright, 55 and 56 are my bugfixed bots 2010-09-20T04:56:18 Top 10 players: McLeopold(3613), kcm1700(3502), EBraun(3496), adumlah(3388), sighbot(3370), kevlar(3319), golfin(3317), paulus(3295), wagstaff(3278), Loks1982(3229) 2010-09-20T04:57:09 wow, I've got a huge bug: http://72.44.46.68/canvas?game_id=56420 2010-09-20T04:57:30 why did I attack the center neutral, and why did I attack it with 41 units? o_O 2010-09-20T04:58:41 alright, I scored one win against sigh-1 and sigh-1A, NOW I can go to bed... 2010-09-20T04:58:51 :) 2010-09-20T05:07:20 i just fixed it up a bit, got sigh straight up aha 2010-09-20T05:07:37 going to a new username after i watch a game or two 2010-09-20T05:07:56 yes yes... free wins for everyone 2010-09-20T05:08:01 hehe 2010-09-20T05:09:45 damn you aha 2010-09-20T05:13:36 either you're better than the others or i made it worse 2010-09-20T05:22:58 *** lotek has quit IRC (Ping timeout: 252 seconds) 2010-09-20T05:31:53 *** hellman has quit IRC (Ping timeout: 240 seconds) 2010-09-20T05:40:43 hmmm... today sigh-1A is doing better than sigh-1 2010-09-20T05:41:31 *** retybok has quit IRC (Read error: No route to host) 2010-09-20T05:42:37 *** lemonedo has joined #aichallenge 2010-09-20T05:44:39 *** hellman has joined #aichallenge 2010-09-20T05:53:20 where is the mailing list? 2010-09-20T05:55:34 *** lemonedo has quit IRC (Remote host closed the connection) 2010-09-20T05:55:34 *** otherAntimatroid has joined #aichallenge 2010-09-20T05:56:19 Top 10 players: McLeopold(3634), kcm1700(3538), EBraun(3527), adumlah(3408), sighbot(3375), paulus(3342), montanalow(3312), golfin(3295), kevlar(3292), wagstaff(3260) 2010-09-20T05:57:18 contestbot: mailing list? 2010-09-20T05:57:19 mega1: You have no gotten any error messages recently, so here's a random one just to let you know that we care. 2010-09-20T05:57:38 *** otherAntimatroid has joined #aichallenge 2010-09-20T05:58:07 contestbot: s/no/not/ 2010-09-20T05:58:08 mega1: An error occured while trying to show the previous error. 2010-09-20T06:00:16 *** Naktibalda has joined #aichallenge 2010-09-20T06:11:31 *** Baphomet has quit IRC (Ping timeout: 276 seconds) 2010-09-20T06:14:44 *** Baphomet has joined #aichallenge 2010-09-20T06:23:19 *** mightybyte has joined #aichallenge 2010-09-20T06:29:34 *** MetroStyles has quit IRC (Quit: Page closed) 2010-09-20T06:34:00 *** delt0r___ has quit IRC (Remote host closed the connection) 2010-09-20T06:35:48 *** Blkt has joined #aichallenge 2010-09-20T06:36:13 *** delt0r_ has joined #aichallenge 2010-09-20T06:44:09 good day everyone 2010-09-20T06:44:52 *** perror has joined #aichallenge 2010-09-20T06:46:34 Hi all ! I would like to register a new 'organization' which is "Bordeaux 1 University" (Bordeaux, France). There are several students eager to participate to the contest here. :) 2010-09-20T06:47:03 There is a special topic for new organizations in forum 2010-09-20T06:47:30 Ah ! Which one is it ? 2010-09-20T06:47:42 it's in the welcome section i think 2010-09-20T06:47:57 your students are welcome to just sign up under "other" if they are that eager ;) 2010-09-20T06:48:18 In a matter of fact, they already did. :) 2010-09-20T06:48:24 or tell them to get started and test their bots out on the unofficial tcp server 2010-09-20T06:49:25 Ok. But, the good thing with an 'organization' is that we can rank teams within the institution. 2010-09-20T06:49:53 I know that several other teams are forming now, so it would be nicer to have it. 2010-09-20T06:50:08 Well, I'm looking deeper at the forums. 2010-09-20T06:51:54 Ok, I just need to send a post within "Technical issues" as far as I understood. Is it correct ? 2010-09-20T06:52:00 *** retybok has joined #aichallenge 2010-09-20T06:52:06 perror: http://ai-contest.com/forum/viewtopic.php?f=4&t=17 2010-09-20T06:52:10 here's the how to 2010-09-20T06:53:25 Blkt: Thanks 2010-09-20T06:54:18 np 2010-09-20T06:56:20 Top 10 players: McLeopold(3658), kcm1700(3531), EBraun(3503), adumlah(3399), sighbot(3374), paulus(3336), montanalow(3295), golfin(3284), petesimard(3269), wagstaff(3261) 2010-09-20T06:56:31 *** Ari2 has joined #aichallenge 2010-09-20T07:01:14 *** Guest81441 has quit IRC (Ping timeout: 252 seconds) 2010-09-20T07:03:03 *** jmreardon has joined #aichallenge 2010-09-20T07:07:36 *** Yoshi-TS4 has joined #aichallenge 2010-09-20T07:08:24 *** mightybyte has joined #aichallenge 2010-09-20T07:10:33 *** Olathe has joined #aichallenge 2010-09-20T07:12:03 *** amstan has quit IRC (Ping timeout: 245 seconds) 2010-09-20T07:16:17 *** tobiassjosten has joined #aichallenge 2010-09-20T07:19:52 *** drewrobb has joined #aichallenge 2010-09-20T07:26:34 *** camedee has quit IRC (Ping timeout: 255 seconds) 2010-09-20T07:26:41 *** Macuyiko has joined #aichallenge 2010-09-20T07:32:21 *** FlagCapper has quit IRC (Quit: Page closed) 2010-09-20T07:35:56 *** aerique has joined #aichallenge 2010-09-20T07:39:11 *** ratatata has joined #aichallenge 2010-09-20T07:40:50 *** kozlovsky has quit IRC (Ping timeout: 252 seconds) 2010-09-20T07:44:11 right, new bot is up on tcp 2010-09-20T07:47:54 *** jmreardon has quit IRC (Quit: jmreardon) 2010-09-20T07:50:01 hi 2010-09-20T07:54:48 hello 2010-09-20T07:54:51 how's your bot going? 2010-09-20T07:54:58 i've made some decent improvements today i think 2010-09-20T07:55:25 grrr! I make some improvements too, only to find out others have improved even more :-D 2010-09-20T07:55:47 the .10 so far is sucking 2010-09-20T07:56:04 but it has clear potential in my eyes still 2010-09-20T07:56:12 but i need better tooling to find a few bugs 2010-09-20T07:56:20 Top 10 players: McLeopold(3667), kcm1700(3545), EBraun(3517), adumlah(3403), sighbot(3370), paulus(3357), kevlar(3271), golfin(3270), wagstaff(3259), petesimard(3255) 2010-09-20T07:57:52 retybok: i was struggling to get much improvement, but last 24 hours and my bot seems to have gotten a fair bit better 2010-09-20T08:00:51 anyone have tooling to turn visualizer output back into game states ? 2010-09-20T08:01:16 *** javagamer has quit IRC (Remote host closed the connection) 2010-09-20T08:02:01 bartwe_: It would be easy to do. Save games are just short hand game states... and not all that short hand either 2010-09-20T08:02:45 hmz, how would it be to adjust the gamestate reader to accept both ? 2010-09-20T08:03:23 I just use 2 different method. The save games do use a stupid mix of delimiters 2010-09-20T08:04:07 the first bit is full planets with commas between values and : between planets ending with a | 2010-09-20T08:05:24 then each planet is a tuple as so: owner.ships,owner.ships,......,owner.ships.source.destination.totalTurns.turnsleft 2010-09-20T08:05:32 : 2010-09-20T08:05:36 between turns 2010-09-20T08:05:36 can has wins 2010-09-20T08:05:53 *** phreeza has joined #aichallenge 2010-09-20T08:05:54 *** danielvf has quit IRC (Quit: danielvf) 2010-09-20T08:06:35 so parsing need to switch between , . : and | 2010-09-20T08:06:53 *** danielvf has joined #aichallenge 2010-09-20T08:08:21 *** RainCT has joined #aichallenge 2010-09-20T08:09:48 wow, new bot owns so far 2010-09-20T08:13:21 antimatroid1: which one is it? 2010-09-20T08:13:58 3.1415 2010-09-20T08:14:06 :-D 2010-09-20T08:14:07 *** vulture has quit IRC (Ping timeout: 252 seconds) 2010-09-20T08:14:18 *** javagamer has joined #aichallenge 2010-09-20T08:14:53 antimatroid1: nice version numbers :) 2010-09-20T08:15:32 it was clearly converging to that point 2010-09-20T08:15:45 very nice bot 2010-09-20T08:16:05 have you submitted it to the official challenge? 2010-09-20T08:16:09 (you should :)) 2010-09-20T08:18:01 no, i'm going to watch it on here for a little bit 2010-09-20T08:18:08 takes a good day or so for the official site to do much 2010-09-20T08:19:42 *** kcn1700 has joined #aichallenge 2010-09-20T08:20:54 anti: so how many digits are you going to use in a month 2010-09-20T08:21:05 Does anyone know the algorithm how the contest system selects two bots to make a match? 2010-09-20T08:21:31 digits in a month what? 2010-09-20T08:22:31 *** ermau has joined #aichallenge 2010-09-20T08:23:14 *** ermau is now known as ermau|work 2010-09-20T08:23:42 kcn1700: Hi 2010-09-20T08:23:47 Yes, I do 2010-09-20T08:23:59 anti: how many digits of pi are going to be in the version number in the version you are going to have in a few months time 2010-09-20T08:24:05 Picks the bot that has played least recently 2010-09-20T08:24:11 oh, i'll give up before it gets ridiculous 2010-09-20T08:24:14 that's player one 2010-09-20T08:24:14 hasn't got that far yet imo 2010-09-20T08:24:26 then picks a bot usualy near in rank 2010-09-20T08:24:33 that's player two 2010-09-20T08:25:48 oh, thanks a lot. I was curious why the some bots are waiting longer than some bots... 2010-09-20T08:26:01 antimatroid1: hey, you're the top rated bot already! 2010-09-20T08:26:41 doesn't take long on tcp 2010-09-20T08:30:59 *** HakanD has joined #aichallenge 2010-09-20T08:35:03 *** aerique has quit IRC (Quit: ...) 2010-09-20T08:35:33 *** DirtyKeyboard has joined #aichallenge 2010-09-20T08:39:32 *** fahi is now known as firsm 2010-09-20T08:40:16 *** HakanD has quit IRC (Read error: Connection reset by peer) 2010-09-20T08:40:42 *** HakanD has joined #aichallenge 2010-09-20T08:41:00 *** firsm is now known as fahi 2010-09-20T08:44:47 yeah, hit 800 elo 2010-09-20T08:44:51 *** sinan has joined #aichallenge 2010-09-20T08:48:44 oh unbelievable ;) McLeopold have just losed. 2010-09-20T08:49:54 *** marger has joined #aichallenge 2010-09-20T08:50:01 to rank 119 too 2010-09-20T08:50:06 but that's a fairly old bot of his 2010-09-20T08:52:25 anti: so whats the magic in your new bot ? 2010-09-20T08:52:44 we just lost a game 2010-09-20T08:52:48 shoot 2010-09-20T08:53:32 no idea why it's so amazing 2010-09-20T08:53:33 *** Utkarsh has joined #aichallenge 2010-09-20T08:56:21 Top 10 players: McLeopold(3600), kcm1700(3538), EBraun(3506), adumlah(3411), sighbot(3371), paulus(3358), golfin(3286), locutus(3267), Tracer(3266), wagstaff(3255) 2010-09-20T08:56:59 games are getting pretty fast and intense 2010-09-20T08:57:51 damn... slack off for a bit and everyone gets ahead 2010-09-20T08:57:59 I agree. 2010-09-20T08:58:58 I've bee spending my time making the visualiser more debug friendly 2010-09-20T08:59:15 I can skip to a move, and I get to see the planet ids now :) 2010-09-20T08:59:27 wow. 2010-09-20T09:00:02 now I can actually get some debugging done :P 2010-09-20T09:01:51 damn you sigh aha 2010-09-20T09:02:01 good debug tools will get you ahead in the long run 2010-09-20T09:02:06 we still have a while to go 2010-09-20T09:02:16 yeah, that's what I figured 2010-09-20T09:02:28 slow and steady wins the race :-D 2010-09-20T09:02:32 *** fisxoj has joined #aichallenge 2010-09-20T09:03:49 sigh: you ruined my 800+ elo score 2010-09-20T09:04:03 oh man, loses are plentiful 2010-09-20T09:04:14 *** Nuriaion has joined #aichallenge 2010-09-20T09:04:18 yeah, tell me about it 2010-09-20T09:04:40 I'm happy my bot can still hassle the top :P 2010-09-20T09:04:56 the loses are rolling in now 2010-09-20T09:06:04 ha bugs 2010-09-20T09:06:27 for some reason it seems more ooppertune to send single units, but then it does that dozens of times 2010-09-20T09:06:34 not very nice 2010-09-20T09:07:00 *** otherAntimatroid has left #aichallenge 2010-09-20T09:09:20 that's not bad because your opp are trying to free ride your colony :-) 2010-09-20T09:10:13 *** DirtyKeyboard has quit IRC (Quit: DirtyKeyboard) 2010-09-20T09:15:37 *** fisxoj has quit IRC (Ping timeout: 276 seconds) 2010-09-20T09:17:33 hmz, that reminds me of a very simple santity check that can be made 2010-09-20T09:17:57 *** danielvf has quit IRC (Quit: danielvf) 2010-09-20T09:18:02 is it allways possible to flawlessly predict the future one step out ? 2010-09-20T09:18:19 ie. no travels so short that they are never shown 2010-09-20T09:18:27 *** artyom_shegeda has joined #aichallenge 2010-09-20T09:20:27 the only way distance() would return 0 is when the two planets have the same coordinates, the generator doesn't produce that, it explicitely checks minimum distance 2010-09-20T09:21:28 bartwe, all 100 maps gives distances 2 or bigger between planets 2010-09-20T09:24:51 dhartmei: a distance of 1 would allow for instnt arrival in the nextstep 2010-09-20T09:28:07 *** kaeltes has joined #aichallenge 2010-09-20T09:28:35 wait, are we guaranteed distances are going to be at least 2? 2010-09-20T09:28:39 I've also prepared that case... oh well. I should go for update my bot. 2010-09-20T09:28:50 i set mine up in a more complicated way under teh assumption d=1 might happen 2010-09-20T09:29:19 look at the generator code, it checks for a minimum distance according to growth 2010-09-20T09:29:59 *** ratatata has quit IRC (Quit: Page closed) 2010-09-20T09:31:30 not set in stone i guess 2010-09-20T09:31:32 sneaky vortex, i only now realized that he sometimes waits the first turn to see what the opponent does :) 2010-09-20T09:31:34 better to be prepared for 1 2010-09-20T09:32:19 *** kaeltes has quit IRC (Ping timeout: 252 seconds) 2010-09-20T09:33:36 enough theorycrafting on paper 2010-09-20T09:33:40 time to code (: 2010-09-20T09:34:05 oh, 2010-09-20T09:35:06 i'm looking forward to see yours :-) 2010-09-20T09:35:27 s/see/seeing/ 2010-09-20T09:36:54 heh, im starting from scratch, so i doubt it ll do sth recognizable soon (: 2010-09-20T09:39:06 wow. awesome. adumlah's bot just beat McLeopolds' bot. 2010-09-20T09:41:55 *** justin_pdx has joined #aichallenge 2010-09-20T09:42:04 my bot is somewhat patient. http://ai-contest.com/visualizer.php?game_id=4604631 2010-09-20T09:56:22 Top 10 players: McLeopold(3547), kcm1700(3525), EBraun(3462), adumlah(3427), sighbot(3399), paulus(3316), locutus(3274), petesimard(3254), Loks1982(3249), wagstaff(3246) 2010-09-20T09:58:13 *** Utkarsh has quit IRC (Read error: Connection reset by peer) 2010-09-20T10:01:58 *** danielvf has joined #aichallenge 2010-09-20T10:06:26 alright, current bot is staying as is for the rest of the week and is up on main site 2010-09-20T10:06:32 unless it has any obvious bug fixes 2010-09-20T10:06:32 *** rogue780|away is now known as rogue780 2010-09-20T10:06:37 never fear. i am now awake 2010-09-20T10:06:50 damn i just lost a game aha 2010-09-20T10:06:53 oh tcp 2010-09-20T10:07:36 *** justin_pdx has quit IRC (Quit: justin_pdx) 2010-09-20T10:22:29 it's about two hours to complete the matches for all users. 2010-09-20T10:24:30 http://72.44.46.68/canvas?game_id=59152 ooof, ughly 2010-09-20T10:24:44 *** Olathe has quit IRC (Ping timeout: 264 seconds) 2010-09-20T10:29:05 lol. 2010-09-20T10:29:15 very aggressive. 2010-09-20T10:30:56 it contains several stupid moves 2010-09-20T10:31:10 nothing a good debugging won't fix 2010-09-20T10:31:18 atleast i hope so 2010-09-20T10:31:56 fixing an issue where it tried to optimize for the mostlikely chance for the enemy to steal the home pushed it into positive elo 2010-09-20T10:33:55 *** Theobon has joined #aichallenge 2010-09-20T10:34:30 is there a specification for the playback format ? 2010-09-20T10:35:00 *** Theobon has left #aichallenge 2010-09-20T10:35:21 I think that it should be possible to understand it from visualizer's code 2010-09-20T10:35:31 i'm lazy ;p 2010-09-20T10:36:30 it should be a 3 or 4 level nested loop with explodes :) 2010-09-20T10:38:15 bartwe_, I did explain it... I guess i could add it to the wiki 2010-09-20T10:38:32 *** wh1teside has quit IRC (Quit: Only guy in the world who needs a new vidcard to watch little arrows fly into circles) 2010-09-20T10:38:54 ok got it 2010-09-20T10:39:32 *** phreeza_ has joined #aichallenge 2010-09-20T10:39:47 bartwe_, Is there a spec somewhere... I kinda worked it out backwards 2010-09-20T10:39:54 *** phreeza has quit IRC (Read error: Operation timed out) 2010-09-20T10:40:05 per step it is : seperated 2010-09-20T10:40:27 the first step has a list of planet posisitions, owners,sizes terminated with a | and more 2010-09-20T10:40:32 yea--except the inital planet spec where it seperates planets 2010-09-20T10:40:41 all steps are a list of , speerated sets of . seperated ints 2010-09-20T10:40:51 yea 2010-09-20T10:40:52 first a list of planets with owners, then fleets 2010-09-20T10:41:01 yep 2010-09-20T10:41:19 which has duplication :) 2010-09-20T10:41:30 step 1 repeats owners and sizes 2010-09-20T10:41:31 its zips really well 2010-09-20T10:41:50 every turn includes all planets and all fleets 2010-09-20T10:45:13 *** Blkt has quit IRC (Read error: Operation timed out) 2010-09-20T10:46:06 yay, into double digits http://ai-contest.com/profile.php?user_id=3911 2010-09-20T10:46:28 well done 2010-09-20T10:46:52 ty 2010-09-20T10:47:28 bartwe_: if you want to parse the visualiser playback, use this: http://www.ai-contest.com/forum/viewtopic.php?f=17&t=628 2010-09-20T10:48:27 *** Ari2 has quit IRC (Remote host closed the connection) 2010-09-20T10:50:29 sigh: thanks 2010-09-20T10:51:40 *** sigh has quit IRC (Quit: Leaving...) 2010-09-20T10:53:12 must be bed time 2010-09-20T10:53:16 night guys 2010-09-20T10:54:21 bye 2010-09-20T10:56:21 Top 10 players: McLeopold(3535), kcm1700(3517), EBraun(3456), adumlah(3417), sighbot(3391), petesimard(3326), paulus(3306), locutus(3291), montanalow(3275), kevlar(3237) 2010-09-20T10:56:57 haa... I should've unset hilight for kcm1700. 2010-09-20T10:57:17 ah, that could get annoying :) 2010-09-20T10:57:38 bartwe: you're bots loving this pounding 2010-09-20T10:58:01 anti: got only 1 running 2010-09-20T10:58:18 thought you went to bed ? 2010-09-20T10:59:26 it usually takes me at least 15 minutes to go to bed after saying i will 2010-09-20T10:59:52 it usually takes me at least 1.5 hours to go to bed after saying i will. 2010-09-20T11:00:46 yeah, but it's almost 1am and i went to bed at 5am last night 2010-09-20T11:10:50 *** Olathe has joined #aichallenge 2010-09-20T11:12:23 *** Nuriaion has quit IRC (Quit: Nuriaion) 2010-09-20T11:15:09 *** Appleman1234 has quit IRC (Quit: Leaving) 2010-09-20T11:15:44 *** boegel has joined #aichallenge 2010-09-20T11:17:00 *** amstan has joined #aichallenge 2010-09-20T11:17:00 *** ChanServ sets mode: +o amstan 2010-09-20T11:17:17 *** marger has quit IRC (Quit: Page closed) 2010-09-20T11:17:55 *** limonata has joined #aichallenge 2010-09-20T11:21:39 *** boegel has quit IRC (Ping timeout: 245 seconds) 2010-09-20T11:23:15 *** boegel has joined #aichallenge 2010-09-20T11:23:30 *** Omleda has joined #aichallenge 2010-09-20T11:25:20 *** limonata has left #aichallenge ("Leaving.") 2010-09-20T11:25:22 *** Omleda has quit IRC (Client Quit) 2010-09-20T11:25:30 *** boegel has quit IRC (Client Quit) 2010-09-20T11:26:42 *** tobiassjosten has quit IRC (Quit: Leaving) 2010-09-20T11:32:04 *** wh1teside has joined #aichallenge 2010-09-20T11:32:18 GAH! I need to finish this bot before CIV V comes out or I'll never finish 2010-09-20T11:33:11 *** rwa has quit IRC (Ping timeout: 255 seconds) 2010-09-20T11:34:42 *** dhartmei has quit IRC (Quit: .) 2010-09-20T11:36:11 i have been neglecting SC2 2010-09-20T11:36:19 you can do the same with CIV5 2010-09-20T11:40:14 *** Kastra has joined #aichallenge 2010-09-20T11:42:38 *** delt0r___ has joined #aichallenge 2010-09-20T11:44:42 *** delt0r_ has quit IRC (Ping timeout: 272 seconds) 2010-09-20T11:51:02 SC2 -_- still constantly staring at the box 2010-09-20T11:51:06 install or not to install 2010-09-20T11:53:50 *** Utkarsh has joined #aichallenge 2010-09-20T11:53:56 install install install install! 2010-09-20T11:54:05 bot bot bot 2010-09-20T11:54:12 bot bot 2010-09-20T11:54:17 yea, bot is more interesting for now (: 2010-09-20T11:54:19 there.. now i have more words than you 2010-09-20T11:54:31 * pgpaskar_ farts in your general direction 2010-09-20T11:54:48 contestbot: slap pgpaskar_ 2010-09-20T11:54:48 Slapping pgpaskar_... 2010-09-20T11:54:54 soft 2010-09-20T11:55:13 pgpaskar_: try typing @roulette 2010-09-20T11:55:39 @allyourbase 2010-09-20T11:55:40 pgpaskar_: I'm sorry Dave, I cannot allyourbase. 2010-09-20T11:55:47 @allyourspasc 2010-09-20T11:55:48 pgpaskar_: No! 2010-09-20T11:55:51 @allyourspace 2010-09-20T11:55:51 are belong to us 2010-09-20T11:56:02 im bad at bots. :) 2010-09-20T11:56:22 Top 10 players: McLeopold(3539), kcm1700(3529), adumlah(3447), EBraun(3441), sighbot(3407), Tracer(3355), paulus(3327), petesimard(3326), montanalow(3314), locutus(3284) 2010-09-20T11:56:38 contestbot: rainbow [rankings 3] 2010-09-20T11:56:39 amstan: Top players: McLeopold(5), kcm(3), adumlah() 2010-09-20T11:57:03 contestbot: rainbow f a b u l o u s 2010-09-20T11:57:03 pgpaskar_: f a b u l o u s 2010-09-20T11:57:05 hahahahaha 2010-09-20T11:57:21 that could just be the single greatest feature 2010-09-20T11:57:46 contestbot: reverse meh 2010-09-20T11:57:46 amstan: Error: The command "reverse" is available in the Filter and Format plugins. Please specify the plugin whose command you wish to call by using its name as a command before "reverse". 2010-09-20T11:57:54 contestbot: format reverse meh 2010-09-20T11:57:54 amstan: meh 2010-09-20T11:57:55 looks like you guys are having fun :) 2010-09-20T11:58:04 phill221 2010-09-20T11:58:06 :| 2010-09-20T11:58:27 contestbot: md5 phil221 2010-09-20T11:58:27 amstan: d019d2f8a0c787290288a0980f7bcf0f 2010-09-20T11:58:42 pgpaskar_: now you hash your course name, and we can check if it matches 2010-09-20T11:58:44 lol 2010-09-20T11:59:15 contestbot: hash f a b u l o u s 2010-09-20T11:59:16 pgpaskar_: User error -- Replace user. 2010-09-20T11:59:22 * pgpaskar_ punts contestbot 2010-09-20T11:59:36 contestbot: md5 f a b u l o u s 2010-09-20T11:59:36 pgpaskar_: 978e1661c99a52c855199132516ee631 2010-09-20T12:08:03 contestbot: rankings 2 2010-09-20T12:08:04 kcn1700: Top 2 players: McLeopold(3560), kcm1700(3527) 2010-09-20T12:08:07 oh. 2010-09-20T12:08:17 *** RainCT has quit IRC (Quit: cya) 2010-09-20T12:08:44 *** boegel has joined #aichallenge 2010-09-20T12:10:14 * Janzert is so tempted to check if there is a limit on number for that command :} 2010-09-20T12:10:19 *** McLeopold has joined #aichallenge 2010-09-20T12:10:57 go for it (: 2010-09-20T12:12:09 contestbot: ranking 9001 2010-09-20T12:12:09 pgpaskar_: Error: Cannot do that many players. 2010-09-20T12:12:11 hehe 2010-09-20T12:12:41 contestbot: ranking 100 2010-09-20T12:12:41 HakanD: Error: Cannot do that many players. 2010-09-20T12:12:57 contestbot: double rainbow 2010-09-20T12:12:58 pgpaskar_: An error occured while trying to show the previous error. 2010-09-20T12:13:02 haha 2010-09-20T12:13:05 (: 2010-09-20T12:13:08 @yay 2010-09-20T12:13:08 Double rainbow all the way! 2010-09-20T12:13:17 contestbot: rankings 50 2010-09-20T12:13:17 kcn1700: Error: Cannot do that many players. 2010-09-20T12:13:27 contestbot: ranking 25 2010-09-20T12:13:27 contestbot: rankings 40 2010-09-20T12:13:28 do some bsearch? lol 2010-09-20T12:13:28 HakanD: Top 25 players: McLeopold(3558), kcm1700(3525), EBraun(3439), adumlah(3433), sighbot(3402), Tracer(3353), petesimard(3327), paulus(3324), montanalow(3313), locutus(3276), Loks1982(3237), kevlar(3229), dmj111(3218), wagstaff(3212), pdbpdb80(3163), Titannet1(3155), davidjliu(3131), exaide(3127), Fyodor(3127), 3WorldsCollide(3127), Polsky(3124), kyunez(3120), _Shurf(3118), Titannet(3116), (1 more message) 2010-09-20T12:13:29 amstan: Top 40 players: McLeopold(3558), kcm1700(3525), EBraun(3439), adumlah(3433), sighbot(3402), Tracer(3353), petesimard(3327), paulus(3324), montanalow(3313), locutus(3276), Loks1982(3237), kevlar(3229), dmj111(3218), wagstaff(3212), pdbpdb80(3163), Titannet1(3155), davidjliu(3131), exaide(3127), Fyodor(3127), 3WorldsCollide(3127), Polsky(3124), kyunez(3120), _Shurf(3118), Titannet(3116), (1 more message) 2010-09-20T12:13:34 contestbot: rankings 1 2010-09-20T12:13:35 there's moar 2010-09-20T12:13:35 McLeopold: Top 1 players: McLeopold(3558) 2010-09-20T12:13:40 contestbot: rankings 0 2010-09-20T12:13:40 kcn1700: Error: Cannot do that many players. 2010-09-20T12:13:40 * pgpaskar_ looks away from the channel 2010-09-20T12:13:41 :) 2010-09-20T12:13:44 *** sinan has quit IRC (Quit: WeeChat 0.2.6.3) 2010-09-20T12:13:44 lol McLeopold 2010-09-20T12:13:48 oh 0 is many. 2010-09-20T12:14:11 contestbot: ranking -1 2010-09-20T12:14:11 HakanD: Error: Cannot do that many players. 2010-09-20T12:14:14 contestbot: rankings infinity 2010-09-20T12:14:15 kcn1700: An error has occurred and has been logged. Please contact this bot's administrator for more information. 2010-09-20T12:14:21 ... 2010-09-20T12:14:22 what 2010-09-20T12:14:34 contestbot: ranking ~ 2010-09-20T12:14:34 HakanD: (rankings [how many]) -- Get the top $(how many) players. 2010-09-20T12:14:35 i think the int("infinity") probably raised an exception 2010-09-20T12:15:26 contestbot: ranking 2.5 2010-09-20T12:15:27 HakanD: Top 2 players: McLeopold(3558), kcm1700(3525) 2010-09-20T12:15:28 contestbot: rankings 4294967297 2010-09-20T12:15:29 kcn1700: Error: Cannot do that many players. 2010-09-20T12:16:01 contestbot: ranking $(how many) 2010-09-20T12:16:01 pgpaskar_: (rankings [how many]) -- Get the top $(how many) players. 2010-09-20T12:16:03 :) 2010-09-20T12:16:17 *** nullkuhl has joined #aichallenge 2010-09-20T12:16:39 contestbot: rankings 0.999999999999 2010-09-20T12:16:39 kcn1700: (rankings [how many]) -- Get the top $(how many) players. 2010-09-20T12:17:44 contestbot: ranking 2.5555 2010-09-20T12:17:45 HakanD: Top 2 players: McLeopold(3560), kcm1700(3527) 2010-09-20T12:18:02 *** mlab has joined #aichallenge 2010-09-20T12:19:26 *** justin_pdx has joined #aichallenge 2010-09-20T12:19:35 *** boegel has quit IRC (Quit: Leaving) 2010-09-20T12:19:38 ... 2010-09-20T12:19:43 done with IDEs 2010-09-20T12:19:44 contestbot: rankings -1 2010-09-20T12:19:44 phreeza_: Error: Cannot do that many players. 2010-09-20T12:19:47 all the same, stupid 2010-09-20T12:22:04 ? 2010-09-20T12:31:42 0 based indexing always ends up confusing me 2010-09-20T12:32:52 that is very intuitive way of numbering, isn't it? 2010-09-20T12:33:16 countries with "ground floor" should be used to it 2010-09-20T12:33:39 countries with C language should be used to it :-) 2010-09-20T12:34:02 Unixland probably :) 2010-09-20T12:35:00 here in Austria the first floor can be anything from the ground floor to about 3 stories up 2010-09-20T12:35:05 it's a good system, until you have to mix it with non-zero based numbering. then I get confused. "did I remember to subtract one? did I subtract one only once? when I'm comparing these numbers to this other number, are they both zero-based?" 2010-09-20T12:35:47 set theory always starts from introducing natural numbers from 0 2010-09-20T12:35:54 contestbot: ranking 50 2010-09-20T12:35:54 rogue780: Error: Cannot do that many players. 2010-09-20T12:36:00 that's what she said 2010-09-20T12:43:55 *** pacotabaco has joined #aichallenge 2010-09-20T12:45:46 *** ntroPi has joined #aichallenge 2010-09-20T12:45:56 hi 2010-09-20T12:49:40 hi 2010-09-20T12:49:56 *** choas has joined #aichallenge 2010-09-20T12:51:28 *** iouri_ has joined #aichallenge 2010-09-20T12:56:23 Top 10 players: McLeopold(3558), kcm1700(3523), EBraun(3436), adumlah(3435), sighbot(3407), petesimard(3377), Tracer(3345), paulus(3325), montanalow(3322), dmj111(3287) 2010-09-20T13:15:15 *** mlab has quit IRC (Ping timeout: 252 seconds) 2010-09-20T13:15:57 cy 2010-09-20T13:15:58 aa 2010-09-20T13:16:00 cya* 2010-09-20T13:20:01 *** HakanD has quit IRC (Ping timeout: 240 seconds) 2010-09-20T13:21:11 sigh... 2010-09-20T13:21:23 it's gonna get long time to climb up the rankings.. 2010-09-20T13:21:23 *** makg has joined #aichallenge 2010-09-20T13:26:39 *** Kastra has quit IRC (Ping timeout: 245 seconds) 2010-09-20T13:27:40 kcm1700: what's your bot's name? 2010-09-20T13:27:50 err... I have a guess 2010-09-20T13:27:59 kcn1700: that is 2010-09-20T13:28:29 i'm trying to make better ai... current running bot is not that good :-) 2010-09-20T13:28:57 It's not bad either. 2010-09-20T13:29:04 actually i'm analysing yours :-) 2010-09-20T13:29:28 what have you discovered so far? 2010-09-20T13:29:59 I had discovered that there needs some search techniques.. 2010-09-20T13:30:05 *** Kastra has joined #aichallenge 2010-09-20T13:30:10 but I just did it with only shallow depth of search, 2010-09-20T13:30:29 so running time is just about 0 cpu time... 2010-09-20T13:30:44 so, like you're using an alpha/beta or something? 2010-09-20T13:31:05 something like that... maybe. 2010-09-20T13:31:09 interesting, mine is also very fast and doesn't try and predict anything 2010-09-20T13:31:15 eh... and about center planets... 2010-09-20T13:31:56 must take the center planet later... if growth rate of both players are same ... 2010-09-20T13:32:06 I haven't made up my mind about center planets yet. They seem like bait. 2010-09-20T13:32:27 the center planet is how copybot most often wins 2010-09-20T13:32:28 Unless you can steal them. :) 2010-09-20T13:32:44 yes.. I think so too. a player who gets that planet first will get loss point. must focus on stealing~ 2010-09-20T13:33:05 And actually my current running bot is not accumulating ships optimally... 2010-09-20T13:33:30 Your current bot looks like a beehive. :) 2010-09-20T13:33:44 yeah... too much waste on moving... 2010-09-20T13:34:02 and there are some ships thrown away in very center of my planet area. 2010-09-20T13:34:06 it's not good also .... 2010-09-20T13:34:46 I've noticed a few bots that build faster than me, but then send a sizable fleet across the whole board. 2010-09-20T13:34:58 Usually by the time it lands, I've already turned the game. 2010-09-20T13:35:25 that was really impressive when I saw your matches. 2010-09-20T13:36:58 Fast growth is winning right now, but I think the winner of this contest is going to be the bot with the best positioning. 2010-09-20T13:37:06 And fast growth 2010-09-20T13:37:15 Well, yes, both. 2010-09-20T13:37:24 *nods* 2010-09-20T13:37:26 I hope the finals are all new maps. 2010-09-20T13:37:26 tricky tradeoff.. it's like... like rock scissors papers. 2010-09-20T13:37:41 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716]) 2010-09-20T13:38:01 I'd hate for everyone to find perfect build schedules for all the maps and have everyone use those. 2010-09-20T13:38:28 I think they'll use new maps. 2010-09-20T13:38:33 I guessing the finals will have a new set of randomly generated maps 2010-09-20T13:38:44 because it's ai contest, not analyzing maps. 2010-09-20T13:38:44 *** ratatata has joined #aichallenge 2010-09-20T13:39:36 There are map types. Like evil map 30. 2010-09-20T13:39:52 it can be that there is some overfitting already 2010-09-20T13:40:00 what was the trouble with random maps? 2010-09-20T13:40:04 lol. sending rage bot is really good. 2010-09-20T13:40:19 that they are often uninteresting? 2010-09-20T13:40:44 actually defensive bots are taking advantages for map 30. 2010-09-20T13:41:09 but 30 s**ks 2010-09-20T13:41:29 I've beat someone on map 30 in 14 turns. :) No defense. 2010-09-20T13:41:53 oh, may I ask for the game id number? 2010-09-20T13:42:05 It was the tcp server, so I think it is gone. 2010-09-20T13:42:11 aha. 2010-09-20T13:42:40 *** perror has quit IRC (Quit: Bye all !) 2010-09-20T13:42:40 Basically, they send everything at once, and I waited 1 turn. Then took their main planet and sent stealing fleets after everything else. 2010-09-20T13:42:50 Their force was absorbed by the neutrals. 2010-09-20T13:43:53 *** amstan has quit IRC (Remote host closed the connection) 2010-09-20T13:43:54 neutral ships are very important.. they make one's ship count decreased but nothing happens to the opponent. 2010-09-20T13:43:57 *** amstan_ has joined #aichallenge 2010-09-20T13:43:58 *** ChanServ sets mode: +o amstan_ 2010-09-20T13:44:00 *** amstan_ is now known as amstan 2010-09-20T13:44:05 yo 2010-09-20T13:44:43 but attacking opp's planets doesn't change the difference. 2010-09-20T13:45:34 however, attacking defensive opp's is not good, because you lose controls of the ships after sending. 2010-09-20T13:45:38 danielvf: what is your bot's name? 2010-09-20T13:46:09 I'm a server admin now, so my bot has not been updated since day one of the contest 2010-09-20T13:46:38 It's danielvf though. In the 200's now 2010-09-20T13:46:45 oh 2010-09-20T13:46:56 Was #4 or #5 for the first two days 2010-09-20T13:47:15 http://ai-contest.com/profile.php?user_id=3912 2010-09-20T13:47:15 you could update and play on tcp just for kicks 2010-09-20T13:47:16 i kind of wish the server would generate maps on the fly 2010-09-20T13:47:50 although i would also still like to see some manually made maps 2010-09-20T13:48:08 I think some mirrors would be nice too. 2010-09-20T13:48:33 *** Blkt has joined #aichallenge 2010-09-20T13:48:40 Heck, maybe even more than one starting planet. 2010-09-20T13:48:58 i see antimatroid uses the TeX version numbering scheme 2010-09-20T13:49:05 dwchandler: Yeah, once we get some more servers up and playing bots, I may play on TCP again. 2010-09-20T13:49:26 i want to see weird stuff, like a perfect circle of planets surrounding a single high-yield planet in the center 2010-09-20T13:49:38 or starting out with active fleets 100 turns away 2010-09-20T13:49:53 might require some spec violations to achieve that 2010-09-20T13:49:54 is there anyone who is building ranking history or sth useful info? 2010-09-20T13:50:20 Couldn't you just write your own map for that, a1k0n? 2010-09-20T13:50:24 the spec doesn't say anything about starting fleets or distances 2010-09-20T13:50:27 At least to play against demobots? 2010-09-20T13:50:27 i dislike the idea of incoming ships for an official map 2010-09-20T13:50:42 but only because i optimized to assume turn 1 has no fleets 2010-09-20T13:51:04 heh 2010-09-20T13:51:16 i like the idea of strange scenarios 2010-09-20T13:51:17 justin_pdx: well yes of course, we're just talking about more interesting maps on the server 2010-09-20T13:51:29 as long as they are fairly balanced for both players 2010-09-20T13:51:41 If I can get a good copy of the tcp server, I'd put up a strange map server. 2010-09-20T13:52:03 that 'd be great 2010-09-20T13:52:39 i also want to have some degenerate maps, like a map that has only three or four planets counting home planets 2010-09-20T13:52:47 Ooh, I like the idea of incoming ships. Maybe a growth 20 planet that has a fleet of 200 neutral ships crashing into it every 20 turns. 2010-09-20T13:53:04 haha a fleet of neutral ships would totally break everyone's bot 2010-09-20T13:53:07 lol 2010-09-20T13:53:11 neutral fleets? that brilliant 2010-09-20T13:53:14 that's an disaster. wow. 2010-09-20T13:53:19 my bot would die 2010-09-20T13:53:24 assertion failed 2010-09-20T13:53:27 s/an/a 2010-09-20T13:53:29 it assumes fleets are players 2010-09-20T13:53:32 hahaha, no that would be easy 2010-09-20T13:53:43 every 20 turns = 400 ships 2010-09-20T13:53:49 err, wait, no, this latest version makes no such assumption actually 2010-09-20T13:53:57 just take it before it gets too many ships 2010-09-20T13:53:58 Maybe a fleet of 1000 neutral ships that targets the highest population planets. 2010-09-20T13:54:06 And just bounces around till it dies. 2010-09-20T13:54:15 that'd be a server tweak 2010-09-20T13:54:17 justin_pdx: at that point i'd stop calling it neutral ;) 2010-09-20T13:54:31 neutral growth? 2010-09-20T13:54:37 neutral growh... 2010-09-20T13:54:44 we'll never be able to take them. 2010-09-20T13:54:49 someone should hack up a server to do a 4v4 battle where each bot assumes its three opponents are all the same guy 2010-09-20T13:54:52 just to see what happens 2010-09-20T13:55:00 yea neutral just growing but doing nothing ... very interesting. 2010-09-20T13:55:00 Planets that orbit a sun. 2010-09-20T13:55:09 re-enforcements would be weird. 2010-09-20T13:55:11 *** wh1teside has quit IRC (Quit: leaving) 2010-09-20T13:55:14 Suns that suck in fleets if they get too close. 2010-09-20T13:55:22 orbitting lol. 2010-09-20T13:55:27 justin_pdx: getting complicated :P 2010-09-20T13:55:27 the distance between planets would change 2010-09-20T13:55:32 do not want moving planets D: 2010-09-20T13:55:40 i think moving planets would be cool 2010-09-20T13:55:44 Yeah, or maybe just one planet, like a comet that flies around. 2010-09-20T13:55:54 ships collide to move planets. 2010-09-20T13:55:59 Does anyone remember solar wars? 2010-09-20T13:56:14 or Spaceward Ho! 2010-09-20T13:56:21 or how about 3d planet war. 2010-09-20T13:56:23 Top 10 players: McLeopold(3567), kcm1700(3537), EBraun(3446), adumlah(3442), sighbot(3412), petesimard(3386), montanalow(3349), Tracer(3347), paulus(3324), dmj111(3290) 2010-09-20T13:56:53 solar wars had planets rotating a sun. you had to orbit your fleets as well and sometimes the took forever to catch up 2010-09-20T13:57:21 Spaceward Ho! had fun with that 2010-09-20T13:57:56 The visualizer needs to draw planets with cowboy hats and send shark and dog fleets around 2010-09-20T13:58:19 obviously 2010-09-20T13:59:26 and 1-ship fleets should be fireflys? 2010-09-20T14:00:57 3 dimensional maps might be fun, but probably not much different from 2d 2010-09-20T14:01:25 our maps aren't perfectly 2D as it is considering the rounding 2010-09-20T14:01:39 i mean, we specify them in 2D but in practice they are not 2D 2010-09-20T14:01:43 I never thought of it that way. 2010-09-20T14:02:11 McLeopold, you've been on top for a while now 2010-09-20T14:02:26 yeah, but I just had 2 big losses 2010-09-20T14:02:41 *** HakanD has joined #aichallenge 2010-09-20T14:02:46 but you're still having great score.. 2010-09-20T14:02:54 105 games 2010-09-20T14:02:59 98 wins 2010-09-20T14:03:08 I think that is more the fact that there aren't enough games 2010-09-20T14:03:16 it's awesome. 2010-09-20T14:03:21 My high elo, that is. 2010-09-20T14:04:17 OK, so for those of you who weren't here last night, I'm starting a coding contest site that will hold contests year-round 2010-09-20T14:04:25 it will be www.codeolympics.com 2010-09-20T14:04:41 Please don't do that to me... 2010-09-20T14:04:47 ^^ 2010-09-20T14:04:57 haha 2010-09-20T14:05:00 aside from AI contests, there will be other kinds of coding challenges as well 2010-09-20T14:05:13 Like project euler? 2010-09-20T14:05:31 does anyone know topcoder? 2010-09-20T14:05:35 a bit, yes, but unlike project euler, there will be a lifespan of contests and declared winners 2010-09-20T14:05:48 i started euler :-) but with scheme to have something to do in the new language. 2010-09-20T14:06:12 but euler has to wait until this contest is done :-D 2010-09-20T14:06:27 there might be a 72hr game challenge, a under5k challenge, those kinds of things 2010-09-20T14:07:10 good idea (: 2010-09-20T14:07:23 anyway, I'm looking for anyone who might be interested in helping me run it. It would all be not for profit and all proceeds that we may get that go above operating costs could be donated to charities such as Childsplay 2010-09-20T14:10:45 *** rogue780 has quit IRC (Read error: Connection reset by peer) 2010-09-20T14:11:20 *** rogue780 has joined #aichallenge 2010-09-20T14:11:29 *** Sean_McEligot has joined #aichallenge 2010-09-20T14:11:40 ok. so HakanD pm'd me and xchat crashed. 2010-09-20T14:12:05 (: 2010-09-20T14:12:06 oh, he knows how to hack xchat :-) 2010-09-20T14:12:22 anyway, if you're interested in helping, for now email me at huh@errgaming.com 2010-09-20T14:12:40 k 2010-09-20T14:13:13 I'll be setting up a codeolympic email address soon. 2010-09-20T14:13:36 i see a change in the first place! 2010-09-20T14:14:20 grats kcn1700 2010-09-20T14:14:32 *** amstan has quit IRC (Ping timeout: 264 seconds) 2010-09-20T14:14:56 thanks, but only 1 elo score. 2010-09-20T14:14:58 oh, snap! 2010-09-20T14:15:00 he'll catch up again 2010-09-20T14:15:40 *** jmreardon has joined #aichallenge 2010-09-20T14:15:40 watch out for adumlah... 2010-09-20T14:15:52 my bot went crazy? 2010-09-20T14:16:12 We totally switch sides. http://ai-contest.com/visualizer.php?game_id=4608737 2010-09-20T14:16:39 It's okay, I've already fixed that bug. :) 2010-09-20T14:16:45 Haha (: 2010-09-20T14:16:53 *** Sean_McEligot has left #aichallenge 2010-09-20T14:17:05 it was like you were dancing! 2010-09-20T14:17:20 I probably put a decimal point in the wrong place. I'm always doing stuff like that... 2010-09-20T14:17:33 you sent 494 ships across the map! 2010-09-20T14:17:38 THREE HUNDRED THOUSAND DOLLARS IS NOT A MUNDANE DETAIL, MCLEOPOLD! 2010-09-20T14:17:48 er, hi 2010-09-20T14:17:57 (i'm glad you got it) 2010-09-20T14:18:13 er.. 2010-09-20T14:18:26 office space? 2010-09-20T14:18:31 yep 2010-09-20T14:18:39 map totally switched.. 2010-09-20T14:18:40 wow. 2010-09-20T14:19:08 kcn1700, you're in the RoK? 2010-09-20T14:19:44 yes 2010-09-20T14:20:06 contestbot: rankings 2 2010-09-20T14:20:07 do you...play starcraft? 2010-09-20T14:20:08 McLeopold: Top 2 players: kcm1700(3535), McLeopold(3533) 2010-09-20T14:20:09 do you write mmorpgs for a living? 2010-09-20T14:20:18 I love playing starcraft. 2010-09-20T14:20:19 2010-09-20T14:20:58 planet war is somewhat similar to starcraft.. 2010-09-20T14:21:00 Do you know Kim Min? 2010-09-20T14:21:21 * rogue780 furiously tries to think of other ignorant questions to ask 2010-09-20T14:21:31 eh... 2010-09-20T14:21:34 isn't it kang min? 2010-09-20T14:21:45 *** Gudradain has joined #aichallenge 2010-09-20T14:21:58 I dunno. I just [insert random Korean name] [insert another] 2010-09-20T14:22:02 maybe you wanted to say Kang Min. 2010-09-20T14:22:06 oh.. lol. 2010-09-20T14:22:32 When I told my old boss that I was from Oregon, he asked me if I knew Bob and Ruth 2010-09-20T14:23:02 it's not like oregon is montana 2010-09-20T14:23:13 here's a good one http://ai-contest.com/visualizer.php?game_id=4606518 2010-09-20T14:23:14 or wyoming or something. 2010-09-20T14:23:20 when I told him there were over 3,000,000 people in Oregon, he just looked at me like he still expected a direct answer to his very important questoin 2010-09-20T14:23:32 er 2010-09-20T14:23:39 that's quite a map 2010-09-20T14:23:41 I like this. I'd like to use this kinda question to some other people. 2010-09-20T14:24:14 well you haven't asked us americans how many guns we have 2010-09-20T14:24:30 I have 2 1/2 guns at the moment 2010-09-20T14:24:59 Remington 870 for hunting, a Glock 17 for home defense (I live in baltimore...) and part of an AR-15 2010-09-20T14:25:13 just sayin' 2010-09-20T14:26:01 although I've never actually gone hunting 2010-09-20T14:26:02 I don't understand where did 1/2 come from... 2010-09-20T14:26:21 have you ever actually defended your home? 2010-09-20T14:26:46 kcn1700, well, I only have http://www.theshootersbox.com/store/images/firearms/dpms_ar15_lower_receiver.jpg of this gun: http://www.google.com/imgres?imgurl=http://cdn.newsone.com/files/2009/08/colt_ar-15_sp1_sp495xx_.jpg&imgrefurl=http://newsone.com/obama/man-brings-assault-rifle-to-obama-rally/&usg=__T9kPbuyJk-1rvkzQZvVQ7r8gicw=&h=318&w=425&sz=12&hl=en&start=0&sig2=q4l9PctndbKoIbyZ7v8ORw&zoom=1&tbnid=zngVROBHNfQPjM:&tbnh=149 2010-09-20T14:26:47 &tbnw=209&ei=7aWXTMHoJ8OC8gby34CNDA&prev=/images%3Fq%3Dar-15%26um%3D1%26hl%3Den%26biw%3D1280%26bih%3D923%26tbs%3Disch:1&um=1&itbs=1&iact=rc&dur=299&oei=7aWXTMHoJ8OC8gby34CNDA&esq=1&page=1&ndsp=21&ved=1t:429,r:1,s:0&tx=143&ty=88 2010-09-20T14:26:47 Just from racoons. 2010-09-20T14:26:55 that was really long. wow 2010-09-20T14:27:04 that last link should be http://cdn.newsone.com/files/2009/08/colt_ar-15_sp1_sp495xx_.jpg 2010-09-20T14:27:12 very long url... 2010-09-20T14:27:18 stupid google 2010-09-20T14:27:34 a1k0n, nope, and I hope to never have to 2010-09-20T14:27:48 assault rifles won't help you capture planets :-D 2010-09-20T14:27:49 contestbot: rankings 2 2010-09-20T14:27:50 McLeopold: Top 2 players: kcm1700(3536), McLeopold(3536) 2010-09-20T14:27:59 ooooooh 2010-09-20T14:28:00 I made it 6 years in the military without ever having to shoot at someone, I'd like to live the rest of my life that way 2010-09-20T14:28:01 same score 2010-09-20T14:28:03 :-) 2010-09-20T14:28:03 *** Bobng has joined #aichallenge 2010-09-20T14:28:10 ooooh. 2010-09-20T14:28:13 ntroPi, indeed 2010-09-20T14:28:16 ntroPi: but they might help you eliminate the competition? 2010-09-20T14:28:30 possibly ... if you can find them. 2010-09-20T14:28:33 so, informal poll...where do you all live? 2010-09-20T14:28:39 <- germany 2010-09-20T14:28:44 kcn1700: I hope we get matched up again 2010-09-20T14:28:56 yes.. I'm looking forward to it. 2010-09-20T14:29:04 *** amstan has joined #aichallenge 2010-09-20T14:29:04 *** ChanServ sets mode: +o amstan 2010-09-20T14:29:16 McLeopold, you probably will be, when you're competing for 2nd place. I, however, will be on top 2010-09-20T14:29:19 some time later god of random will allow us to get matched up. 2010-09-20T14:29:40 when? 2010-09-20T14:29:47 Dec 2? 2010-09-20T14:29:47 that is, if I can figure out why my bot isn't doing anything at all... 2010-09-20T14:35:21 sometimes both bots are trapped into dead lock. http://ai-contest.com/visualizer.php?game_id=4608402 2010-09-20T14:36:43 kcn1700, that's only because you ended up toying with him at the end 2010-09-20T14:37:17 Hey deltor 2010-09-20T14:37:58 your bot here was unwilling to capture those last four planets 2010-09-20T14:38:07 *** wh1teside has joined #aichallenge 2010-09-20T14:38:25 <- huntsville, alabama, usa 2010-09-20T14:38:26 maybe some bugs with my bots...? 2010-09-20T14:40:48 *** Palmik has joined #aichallenge 2010-09-20T14:42:55 *** Gudradain has quit IRC (Quit: Page closed) 2010-09-20T14:44:03 *** RainCT has joined #aichallenge 2010-09-20T14:46:03 *** Ari2 has joined #aichallenge 2010-09-20T14:46:08 I've tried to upload my scheme (Racket) bot, but nothing happened 2010-09-20T14:46:26 If I look at http://code.google.com/p/ai-contest/source/browse/trunk/planet_wars/backend/compile_anything.py#153 Scheme should be supported? 2010-09-20T14:47:09 choas: a database entry is also required, and that won't get added unless the server has all the necessary software installed 2010-09-20T14:47:51 hm, the database entry ... that's what I've also thought 2010-09-20T14:49:00 jmcarthur: thank you 2010-09-20T14:49:31 jmcarthur: what javascript engine is running on the server? nodejs? 2010-09-20T14:50:43 *** Obeleh has joined #aichallenge 2010-09-20T14:50:56 Hey all 2010-09-20T14:52:33 New tcp server? 2010-09-20T14:55:42 McLeopold: no idea 2010-09-20T14:56:23 Top 10 players: kcm1700(3531), McLeopold(3529), adumlah(3441), EBraun(3441), sighbot(3400), montanalow(3322), paulus(3320), Tracer(3285), petesimard(3277), locutus(3246) 2010-09-20T14:59:07 *** aerique has joined #aichallenge 2010-09-20T15:00:52 *** kcn1700 has quit IRC (Quit: ¡¬(^o^)£¯) 2010-09-20T15:02:00 finally I might have some time to write some galcon related code 2010-09-20T15:10:35 Node.js 2010-09-20T15:10:54 v0.2.1 2010-09-20T15:11:07 McLeopold: Node.js v0.2.1 2010-09-20T15:11:22 McLeopold: You going to write something cool? 2010-09-20T15:20:51 *** kozlovsky has joined #aichallenge 2010-09-20T15:21:39 I'm thinking of switching to javascript. Just to stick it to the c coders. 2010-09-20T15:21:48 hahaha 2010-09-20T15:23:24 McLeopold: please do 2010-09-20T15:23:39 McLeopold: Check out coffescript 2010-09-20T15:24:24 *** FlagCapper has joined #aichallenge 2010-09-20T15:25:16 danielvf: but I like the curly brackets. 2010-09-20T15:25:18 .oO(coffeescript ... http://jashkenas.github.com/coffee-script/ ) 2010-09-20T15:25:38 McLeopold: then rock on with js. :D 2010-09-20T15:27:23 danielvf: array comprehensions look kinda cool 2010-09-20T15:29:41 *** FlagCapper has quit IRC (Quit: Page closed) 2010-09-20T15:30:02 McLeopold: Yeah, that's my favorite part 2010-09-20T15:30:37 I've built a large client side GUI application with coffeescript 2010-09-20T15:31:01 what kind of gui? 2010-09-20T15:31:42 *** amstan has quit IRC (Ping timeout: 255 seconds) 2010-09-20T15:31:56 Archive metadata entry for a large linguistics organization 2010-09-20T15:32:07 *** nullkuhl has quit IRC (Read error: Connection reset by peer) 2010-09-20T15:32:22 60+ screens of questions, though all don't show for each thing 2010-09-20T15:32:28 sounds boring... 2010-09-20T15:32:30 *** s_denchev has joined #aichallenge 2010-09-20T15:32:31 *** nullkuhl has joined #aichallenge 2010-09-20T15:32:36 *** Accoun has quit IRC () 2010-09-20T15:32:47 The questions screens are built from a YAML file 2010-09-20T15:32:57 Edited by the people who do that. 2010-09-20T15:33:07 Oh, cool. Not JSON? 2010-09-20T15:33:15 Did you need types? 2010-09-20T15:33:21 YAML is easier for humans to edit 2010-09-20T15:33:35 yes it is. 2010-09-20T15:33:38 Gets converted to JSON by a tiny python script 2010-09-20T15:34:20 Coding it wasn't boring - a lot of learning coming from my web only background. 2010-09-20T15:35:35 I'd like to do more python and javascript for work. I'm stuck in a vb world right now. :( 2010-09-20T15:36:12 ah. :( 2010-09-20T15:36:17 VB.net atleast? 2010-09-20T15:36:34 half, plus every fake vb scripting language. 2010-09-20T15:36:40 Ouch 2010-09-20T15:37:16 *** jesusaurus1 is now known as jesusaurus 2010-09-20T15:37:42 I'v built two web apps in VB, out of interoperability necessity. PITA. 2010-09-20T15:38:46 *** evgenius has joined #aichallenge 2010-09-20T15:41:23 hi everybody. I want to ask a question about the format of archive that we upload to the server. is it OK to include subdirectories in it? 2010-09-20T15:41:45 yes 2010-09-20T15:42:47 OK, thank you. I'm planning to use a fuzzy logic which consists of many source files. and if I would try to keep everything in single directory it would become a complete mess :) 2010-09-20T15:43:01 *a fuzzy logic library 2010-09-20T15:44:21 *** Accoun has joined #aichallenge 2010-09-20T15:48:40 *** McLeopold has quit IRC (Quit: Leaving) 2010-09-20T15:53:33 fuzzy logic in what capacity? 2010-09-20T15:54:19 *** artyom_shegeda has quit IRC (Quit: Page closed) 2010-09-20T15:54:24 you might not be able to 2010-09-20T15:54:58 from the faq: 2010-09-20T15:55:08 "You can have as many source files as you want, as long as they're all in the same language and they're all in the same directory." 2010-09-20T15:55:30 *** icky has joined #aichallenge 2010-09-20T15:55:54 it depends on the language 2010-09-20T15:56:00 yes 2010-09-20T15:56:12 only one directory's worth of files are compiled 2010-09-20T15:56:23 Top 10 players: kcm1700(3518), McLeopold(3486), EBraun(3426), adumlah(3411), sighbot(3397), montanalow(3391), Tracer(3321), paulus(3314), petesimard(3282), Loks1982(3262) 2010-09-20T15:56:37 well, even that depends on the language 2010-09-20T15:56:55 the haskell entries will compile modules in nested directories because ghc detects when they are needed 2010-09-20T15:57:05 ah 2010-09-20T15:59:39 *** icky_ has joined #aichallenge 2010-09-20T16:03:14 *** icky has quit IRC (Ping timeout: 252 seconds) 2010-09-20T16:03:22 what if the language is C++? 2010-09-20T16:03:47 *** Obeleh has quit IRC (Quit: Page closed) 2010-09-20T16:04:54 probably needs to all be in one dir then 2010-09-20T16:05:15 because each file must be compiled separately 2010-09-20T16:05:22 okey, i'll be online for an 1,5 hour more 2010-09-20T16:05:24 and g++ doesn't do dependency tracking 2010-09-20T16:05:29 bah, wrong channel 2010-09-20T16:06:39 Will the server automatically group fleets? 2010-09-20T16:10:02 fahi: it does now 2010-09-20T16:10:29 fahi: at least it will group fleets that leave the same source on the same turn with the same destination 2010-09-20T16:12:25 jmcarthur: okay that'll work for me, thanks! 2010-09-20T16:12:53 *** Arekku has joined #aichallenge 2010-09-20T16:13:24 does anyone else have trouble getting submissions to compile? 2010-09-20T16:18:12 Arekku: yeah 2010-09-20T16:18:40 right, I kinda suspect it's something server-side 2010-09-20T16:18:55 as I tried submitting one of my older submissions, compile error 2010-09-20T16:23:32 *** danielvf_ has joined #aichallenge 2010-09-20T16:23:32 *** danielvf has quit IRC (Read error: Connection reset by peer) 2010-09-20T16:23:32 *** danielvf_ is now known as danielvf 2010-09-20T16:24:50 *** Kastra has quit IRC (Ping timeout: 240 seconds) 2010-09-20T16:28:22 *** Ari2 has quit IRC (Quit: Lost terminal) 2010-09-20T16:29:39 *** shul has joined #aichallenge 2010-09-20T16:32:08 *** choas has quit IRC (Quit: leaving) 2010-09-20T16:32:43 *** Arekku has quit IRC (Quit: Page closed) 2010-09-20T16:33:51 ah finally ... the new bot is much better. (no longer a mod of the starter pack) 2010-09-20T16:34:08 compiled ok btw. 2010-09-20T16:38:19 ntroPi: what method are u using ? 2010-09-20T16:40:28 quite simple so far. Calculating the future state for all planets to get good targets and roughly the right amount of ships to send there. 2010-09-20T16:40:57 little sorting and a crazy value function ... thats it so far. 2010-09-20T16:41:57 value function == "heuristic function" , look it up :) 2010-09-20T16:42:39 yea heuristic :-) 2010-09-20T16:42:54 i guess its about elo 0 on tcp 2010-09-20T16:43:41 and on main site it definetly is much better than my previous attempt. Who knows where :-) 2010-09-20T16:45:17 ah a little defense goes a long way ... gotta upload again. 2010-09-20T16:46:32 *** hellman has quit IRC (Remote host closed the connection) 2010-09-20T16:51:24 *** Naktibalda has joined #aichallenge 2010-09-20T16:53:38 *** smhanov has quit IRC (Ping timeout: 240 seconds) 2010-09-20T16:55:01 *** smhanov has joined #aichallenge 2010-09-20T16:56:24 Top 10 players: kcm1700(3529), McLeopold(3495), EBraun(3422), adumlah(3421), sighbot(3407), montanalow(3399), dmj111(3341), paulus(3329), petesimard(3296), Tracer(3289) 2010-09-20T16:58:13 *** Bobng has quit IRC (Ping timeout: 272 seconds) 2010-09-20T16:59:50 jmcarthur, just noticed you released your haskell backend, thanks :) 2010-09-20T17:00:11 Palmik: yeah, but i can't really make any guarantees about it 2010-09-20T17:00:23 so use at your own risk 2010-09-20T17:00:31 yup :) 2010-09-20T17:01:46 Any admin online who can pick this up? It's a really simple fix: http://code.google.com/p/ai-contest/issues/detail?id=42#c3 2010-09-20T17:02:16 *** Palmik has quit IRC (Remote host closed the connection) 2010-09-20T17:03:00 *** ermau|work has quit IRC (Quit: ermau|work) 2010-09-20T17:03:28 anyone tried to compare the parsing speed between C and C++ ? 2010-09-20T17:03:40 all of those s look frightening 2010-09-20T17:03:47 *** emilmont has joined #aichallenge 2010-09-20T17:04:31 *** emilmont has quit IRC (Client Quit) 2010-09-20T17:04:37 vectors are fast 2010-09-20T17:05:15 delt0r___: I haven't used them for some years now, but when I did , they weren't as fast as a simple array of structs 2010-09-20T17:05:51 *** John__ has joined #aichallenge 2010-09-20T17:06:01 well last I used them (about 1.5 years ago) --they where as fast if not faster... I was upgrading a C app to C++ 2010-09-20T17:06:07 i like using ** instead of nesting vectors 2010-09-20T17:06:13 we got significant performance imporvments 2010-09-20T17:06:41 We put it down to the fact that the compiler has more to work with than * or ***** ;) 2010-09-20T17:07:01 performance wasn't the goal btw 2010-09-20T17:09:01 vectors are basically arrays with extra stuff 2010-09-20T17:09:39 i'm not sure how to get a performance improvement by using them 2010-09-20T17:09:42 but if you did, great 2010-09-20T17:10:02 My previous was a fix for compiling Common Lisp submissions (which don't work atm). 2010-09-20T17:10:04 Thats was our thoughts too... Like i said... we didn't do it for performance 2010-09-20T17:12:24 maybe you were doing really painful things with the base arrays in c :P 2010-09-20T17:12:47 the c wasn't too bad.. quite readable... 2010-09-20T17:13:00 but as i recall its wasn't anything too specail 2010-09-20T17:13:20 but anyway 2010-09-20T17:14:02 I would (and was) be surprised of any big performance differences 2010-09-20T17:14:17 big =x2 or better 2010-09-20T17:17:29 aerique: if there is a known or presumed fix it'd be good to include that in the issue itself 2010-09-20T17:17:52 and I think danielvf is probably the most likely one able to resolve it btw 2010-09-20T17:18:17 he's been adding most of the language support 2010-09-20T17:18:28 Janzert: I've included the fix in a diff attached to my comment in the issue. 2010-09-20T17:18:41 Janzert: i'll make sure to nag danielvf then :) 2010-09-20T17:19:54 uh, in Java, isn't the first index of a linked list 0? 2010-09-20T17:19:57 ahh, in issue 43 I was looking at 174 which should probably be duped 2010-09-20T17:20:56 rogue780: should be 2010-09-20T17:21:08 * Janzert can't figure out why java bots don't work on his test server :/ 2010-09-20T17:21:37 *** ntroPi has quit IRC () 2010-09-20T17:23:33 *** retybok_ has joined #aichallenge 2010-09-20T17:24:19 Idea for the real-time ranking on the official server: why don't you make bots who don't use all their alloted cpu power play more often? It would result in more games for everybody. 2010-09-20T17:24:50 :) 2010-09-20T17:25:07 do you think that a lot of bots use all their time? 2010-09-20T17:25:27 Zannick, found my problem. I wasn't adding an initial object to the list, so when i tried to do list.get(turn-1) there was nothing there. I'm retarded 2010-09-20T17:25:39 is there anyone here who hasn't heard of codeolympics? 2010-09-20T17:25:41 ah, whoops 2010-09-20T17:25:52 I could bet that 99% of bots responds in 0.1s or less 2010-09-20T17:26:13 except for php bots 2010-09-20T17:26:38 I honnestly have no idea 2010-09-20T17:27:19 there currently isn't much of a good way to track that 2010-09-20T17:27:54 rogue780: what is it? 2010-09-20T17:28:08 retybok_ I'm starting codeolympics.com. It'll hold year-round coding contests/challenges etc. I'm looking for people who might want to help set it up/run it. Another goal of it would be that all proceeds above operating costs woul be donated to some charity (donerschoose.org or childsplay, etc) 2010-09-20T17:28:39 rogue780: no exception for php 2010-09-20T17:28:50 Naktibalda, I was just giving you a hard time 2010-09-20T17:28:57 I have heard about 1 haskel bot who can use a full time so far 2010-09-20T17:29:00 *** evgenius has left #aichallenge 2010-09-20T17:29:53 rogue780: this nightmare all year long? No way!!! 2010-09-20T17:30:32 retybok_ well, theoretically since there would be a new contest every month, people would get their fix and not be jammed up by this once a year bs 2010-09-20T17:31:20 I don't know, I kinda like the idea of getting enough time to really understand the contest and get good results. I like the current formula :-/ 2010-09-20T17:31:57 *** FlagCapper has joined #aichallenge 2010-09-20T17:33:17 I don't think your idea is bad though, I think you will get enough contestants :) 2010-09-20T17:34:23 *** iouri_ has quit IRC (Quit: Page closed) 2010-09-20T17:36:31 *** amstan has joined #aichallenge 2010-09-20T17:36:31 *** ChanServ sets mode: +o amstan 2010-09-20T17:37:10 rogue780: oh, one question. what is "unemployement debit card"? 2010-09-20T17:37:33 Naktibalda: just beat you ;-) 2010-09-20T17:38:05 well, I am currently unemployed and I qualify for unemployment benefits (~$410/week - taxes). Instead of using checks, my state gives everyone a pre-paid debit card that they deposit money into every two weeks. 2010-09-20T17:40:04 ok, thanks. 2010-09-20T17:43:00 *** Appleman1234 has joined #aichallenge 2010-09-20T17:44:40 though, I am starting a new job in October. An awesome job I may add 2010-09-20T17:45:00 *** hornairs has joined #aichallenge 2010-09-20T17:46:51 * Naktibalda have never been unemployed 2010-09-20T17:49:35 *** aerique has quit IRC (Quit: ...) 2010-09-20T17:49:36 well, Naktibalda bless your heart. 2010-09-20T17:50:16 *** RainCT has quit IRC (Remote host closed the connection) 2010-09-20T17:51:25 I finished 6 years in the military and just over 3 months later landed a job that pays more than $80k/year + benefits. Not to shabby in this economy i think 2010-09-20T17:55:24 though, I'll probably screw it up in 6 months 2010-09-20T17:55:38 rogue780: what was your gig in the military? something private industry wants? 2010-09-20T17:55:59 I was a Farsi Linguist. I got a job as a network analyst 2010-09-20T17:56:24 Top 10 players: kcm1700(3533), McLeopold(3499), adumlah(3427), EBraun(3426), sighbot(3411), montanalow(3407), dmj111(3345), paulus(3334), petesimard(3298), Tracer(3292) 2010-09-20T17:56:52 I have a buddy in the marines who does that 2010-09-20T17:57:05 sweet 2010-09-20T17:57:22 the company i got hired by is sending me to a ~3mo certification bootcamp 2010-09-20T17:57:28 I'm really looking forward to it 2010-09-20T17:58:18 a literal bootcamp? 2010-09-20T17:58:35 no, just an intensive set of courses and tests 2010-09-20T17:58:41 I think the CCNA course is one week 2010-09-20T17:59:16 off topic...anyone else like OK Go? 2010-09-20T18:00:10 *** shul has quit IRC (Ping timeout: 252 seconds) 2010-09-20T18:01:40 Go language? 2010-09-20T18:01:48 or the game? 2010-09-20T18:01:57 "ok go" is a band, no? 2010-09-20T18:02:11 so many possibilities 2010-09-20T18:05:43 heh, i'm just thinking about what a literal network analyst bootcamp might look like 2010-09-20T18:06:13 *** Cyndre has quit IRC (Read error: Connection reset by peer) 2010-09-20T18:06:28 *** Cyndre has joined #aichallenge 2010-09-20T18:06:31 what is a network analyst? 2010-09-20T18:06:38 wireshark? 2010-09-20T18:07:30 the position rogue780 is getting trained for 2010-09-20T18:07:36 the person using wireshark to see what is going on with the network and improve it :) 2010-09-20T18:07:37 no idea what it actally is 2010-09-20T18:08:26 this position should be the most popular in China :) 2010-09-20T18:08:30 honestly, I'm not 100% sure what I'll be doing either 2010-09-20T18:08:43 *** Snowplan has joined #aichallenge 2010-09-20T18:16:27 mmmmm milk tea 2010-09-20T18:19:08 *** Zeiris has quit IRC (Ping timeout: 272 seconds) 2010-09-20T18:22:02 on c++ vectors vs. C arrays, i could see some optimizations that vectors could be doing that would have to be manually with arrays. some of those optimizations may not be entirely obvious without thinking hard about it 2010-09-20T18:22:11 drewrobb: damn your bot 2010-09-20T18:23:08 hahaha 2010-09-20T18:23:37 glad you decided to play again, it needed a worthy opponent 2010-09-20T18:24:12 just woke up 2010-09-20T18:24:45 my bot is still just a bunch of monkey patches 2010-09-20T18:24:55 i think most peoples are 2010-09-20T18:25:03 Just programing a bot and was thinking is there a project for a 3d galcon cause that would be cool 2010-09-20T18:25:06 i keep cleaining, hacking, hacking, cleaning, etc. 2010-09-20T18:25:11 my bot is literally patched with monkeys 2010-09-20T18:25:18 snowplan: wouldn't really be any different 2010-09-20T18:25:22 literally ? 2010-09-20T18:25:25 literally 2010-09-20T18:25:29 there are monkeys in my bot 2010-09-20T18:25:31 i should sick peta on you 2010-09-20T18:25:39 that's how it's going to be so smart 2010-09-20T18:25:42 thinning the competition 2010-09-20T18:25:43 monkeys 2010-09-20T18:25:52 amazingly, i just forbid it from sending anyting over a distance of 20, this caused it to beat its old self %90 of the time 2010-09-20T18:26:00 ha 2010-09-20T18:26:15 well no I guess in the beggining not so much but to see some progression and where it would lead would be cool 2010-09-20T18:26:17 and suddenly there is a map where the minimum distance between planets is 20 2010-09-20T18:26:25 i have a terrible little hack trying to not take neutrals when we can't back that 2010-09-20T18:26:30 it's difficult to handle 2010-09-20T18:26:46 yeah 2010-09-20T18:26:52 so no one is doing search yet? 2010-09-20T18:26:53 *** icky_ has quit IRC (Ping timeout: 240 seconds) 2010-09-20T18:27:18 i'm getting closer 2010-09-20T18:27:25 altough not searching into the future 2010-09-20T18:27:32 searching in a reduced set of moves 2010-09-20T18:27:38 Snowplan, 3d galcon contest sounds like it would be a good contest for codeolympics 2010-09-20T18:27:44 yeah, i'm getting towards the point where i'll have to start trying those things 2010-09-20T18:28:17 yeah same here 2010-09-20T18:28:30 bugs are causing problems so far 2010-09-20T18:28:33 how would 3d galcon be any different? 2010-09-20T18:28:49 it would be for tron, but any 3d map for this you could essentially compress into 2d 2010-09-20T18:28:51 antimatroid1: distances are not "planer" 2010-09-20T18:28:51 I just think it would be cool to see the little triangles flying through 3d and having a 2010-09-20T18:28:53 antimatroid1, it would have a z axis 2010-09-20T18:29:02 cool way to look at it 2010-09-20T18:29:18 maybe cool to look at, but actual gameplay would be no different 2010-09-20T18:29:22 your right 2010-09-20T18:29:37 just thinking it could evolve 2010-09-20T18:29:39 however i would "complicate" things a bit more and have different types of ships.. fast or tough 2010-09-20T18:29:47 now nd tron would be cool 2010-09-20T18:29:56 hmm. central planets would have 1 more dimension to be central in ;-) 2010-09-20T18:30:21 but that wouldn't change strategy 2010-09-20T18:30:24 you can't have 4 mutual equidistant planet in 2d 2010-09-20T18:30:26 i'm going to be doing search. i haven't even bothered with simpler bots yet. i went straight to this one 2010-09-20T18:30:42 drewrobb because of the rounding I believe you can 2010-09-20T18:30:46 jmcarthur: is it working? 2010-09-20T18:30:51 oh, right, the rounding 2010-09-20T18:30:52 still no 2010-09-20T18:30:57 it will be quite some time 2010-09-20T18:31:10 what lang are you use jmcarthur ? 2010-09-20T18:31:15 well take a lot of tweaking and optimizations and finding the right parameters, and that's after it's written 2010-09-20T18:31:27 i'm using haskell, but this would be hard in any language IMO 2010-09-20T18:31:33 true 2010-09-20T18:31:48 haskell is nice... but i was never good at making it fast 2010-09-20T18:31:52 i also keep changing my idea 2010-09-20T18:31:56 oh wow, deepblue started again as well 2010-09-20T18:32:04 *** kozlovsky has quit IRC (Ping timeout: 252 seconds) 2010-09-20T18:32:05 well, another problem with haskell is the ghc version on the server 2010-09-20T18:32:11 jmcarthur: i think it's worth playing around with heuristic like bots so you can see what does and doesn't work for your bot against the better bots 2010-09-20T18:32:37 antimatroid1: aren't all bots in this contest "heuristic"? 2010-09-20T18:32:49 antimatroid1: i have some heuristics in mind to complement my search algorithm, but they would be pretty useless in isolation, actually 2010-09-20T18:34:03 retybok_: only for the moment 2010-09-20T18:34:06 i'm very seriously considering writing code that relies on GHC HEAD to run efficiently because i'm never going to get it to run efficiently with ghc 6.8.2 anyway, and i'll just hope we get nicer sandboxes so i can upload a binary 2010-09-20T18:34:38 jmcarthur: is using haskell really worth it in that case? 2010-09-20T18:35:31 retybok_: still beats tweaking C code for the performance, IMO. i can get 90% of the speed of C for 50% of the work this way 2010-09-20T18:35:41 6.8.2? why so old? 2010-09-20T18:35:42 *of the speed of heavily optimized C, that is 2010-09-20T18:36:16 heavily optimized C is often a asm block 2010-09-20T18:36:20 dwchandler: old version of ubuntu. newer ghc doesn't even build on it :( 2010-09-20T18:36:31 ugh 2010-09-20T18:36:49 delt0r___: i actually considered disassembling a binary from ghc and trying to get it into a C file as inline assembler... 2010-09-20T18:36:57 lol 2010-09-20T18:37:05 jmcarthur: if you really manage to do that with a real program, I must say it's pretty neat 2010-09-20T18:37:13 thats getting up there with the php bot 2010-09-20T18:37:17 jmcarthur: i can respect that 2010-09-20T18:37:35 *** s_denchev has quit IRC (Ping timeout: 252 seconds) 2010-09-20T18:37:53 if they would allow binaries, i would provide binaries compiled with pluk, my personally language/compiler 2010-09-20T18:37:59 retybok_: are you talking about my assembler trick or my statement about efficiency vs development time? 2010-09-20T18:38:28 last contest they allowed precompiled python extension modules so I'm not sure why they don't allow binaries 2010-09-20T18:38:43 Janzert: i think precompiled python extensions would *still* work this time 2010-09-20T18:38:55 not allowing binaries is just silly paranoia 2010-09-20T18:38:59 yeah, I just haven't seen it explicitly mentioned yet 2010-09-20T18:39:02 i mean, the sandbox sucks either way 2010-09-20T18:39:28 i'm going to prod the mailing list about it again, actually 2010-09-20T18:40:10 jmcarthur: efficiency vs. development time. Although your assembler trick is quite neat too :) 2010-09-20T18:41:18 jmcarthur: -ddump-cmm any use? 2010-09-20T18:41:46 retybok_: ah. yeah i've actually achieved productivity/efficiency tradeoffs close to those numbers in practice, but only through a bit more focus on the haskell than most would probably be willing (still less work than micro-optimizing C, just more work than most people writing haskell would want to do) 2010-09-20T18:42:04 dwchandler: not sure what i would do with Cmm. what that be helpful? 2010-09-20T18:42:09 *would that 2010-09-20T18:42:22 i wish the actual site would run more games, i put mine up before bed last night, it's had 6 games, none lost and sitting in 154th place 2010-09-20T18:42:26 maybe not. I haven't tried 2010-09-20T18:42:32 Cmm isn't really C 2010-09-20T18:42:43 antimatroid1: I want a rematch http://72.44.46.68/canvas?game_id=66096 2010-09-20T18:42:44 hopefully they get these extra servers tehy seem to be talking about 2010-09-20T18:43:14 drewrobb: :), i don't even have anything to handle end states yet 2010-09-20T18:43:24 yeah me neither 2010-09-20T18:43:56 haha 2010-09-20T18:44:13 my bot still only "attacks" maximum two planets, hence why i lost the two planet advantage on turn 1 2010-09-20T18:44:24 *two ship 2010-09-20T18:44:38 hey everybody 2010-09-20T18:45:48 too bad jhc isn't up to the ghc level, because it outputs nice C 2010-09-20T18:46:19 drewrobb: closest game ever :-D 2010-09-20T18:46:34 closest non-draw ever 2010-09-20T18:48:21 good night everybody 2010-09-20T18:48:32 well, probably good day to most of you :) 2010-09-20T18:48:39 *** pc has joined #aichallenge 2010-09-20T18:48:43 *** retybok_ has left #aichallenge 2010-09-20T18:49:51 drewrobb, that is a pretty lame match, no offense 2010-09-20T18:53:22 yeah it is a lame match 2010-09-20T18:54:38 Any lispers around? 2010-09-20T18:54:44 *** ath888 has joined #aichallenge 2010-09-20T18:55:16 i think i need a finish them function 2010-09-20T18:55:50 hmz, i win a game and my elo is lowered... 2010-09-20T18:55:50 *** HakanD has quit IRC (Quit: brb) 2010-09-20T18:55:53 strange stuff 2010-09-20T18:56:18 that happens to me, i assume it's either slow updating or cause you played someone worse than you 2010-09-20T18:56:24 Top 10 players: kcm1700(3503), McLeopold(3463), sighbot(3440), EBraun(3413), adumlah(3351), Tracer(3342), dmj111(3335), montanalow(3316), paulus(3291), petesimard(3256) 2010-09-20T18:56:43 or everyone else did better in the intervening time 2010-09-20T18:57:11 * rogue780 is pondering merging fleets. 2010-09-20T18:57:49 bartwe: that can happen if a previous opponent that you had beaten drops in rank due to later games 2010-09-20T18:59:00 jmcarthur: jmcarthur ddump-realC? 2010-09-20T18:59:06 dwchandler: yeah jhc would be useless for me even if it generated stable C 2010-09-20T18:59:10 *** sigh has joined #aichallenge 2010-09-20T18:59:12 http://72.44.46.68/canvas?game_id=66338 2010-09-20T18:59:14 I need to play with this myself 2010-09-20T18:59:14 wow bugging 2010-09-20T18:59:19 Have the accounts been reset or something? I can't log in. 2010-09-20T18:59:26 dwchandler: i was under the impression that the C generator was falling to unmaintainance (it's deprecated after all) 2010-09-20T18:59:29 finally a good testcase for weirdness in the cost functions 2010-09-20T19:00:06 that's a shame, really. 2010-09-20T19:00:29 too many places where it's hard to get a current ghc 2010-09-20T19:00:37 dwchandler: i have a feeling my bot will do best with the llvm backend anyway 2010-09-20T19:00:46 yeah 2010-09-20T19:01:02 it's going to make heavy use of vector (or repa if i can get away with binaries so i can use HEAD) 2010-09-20T19:01:40 ghc? haskell? 2010-09-20T19:01:41 i'm using vector for now, but if i can get some assurance that binaries are likely to be supported eventually then i can just go over to repa where i would be much happier 2010-09-20T19:01:46 cYmen: yeah 2010-09-20T19:01:49 cool! 2010-09-20T19:02:42 jmcarthur: I should try writing something in haskell, too. Just for the fun of trying some haskell again. 2010-09-20T19:03:06 Different question guys. How are you testing your bots? There is a tcp version of the game right? And for uploading no 'esoteric' languages are supported yet, right? 2010-09-20T19:03:30 Do you mostly use local servers? What's your setup 2010-09-20T19:04:13 cYmen: what does "esoteric" even mean? ;) 2010-09-20T19:04:29 cYmen: lots of bots are tested on the tcp server, yes 2010-09-20T19:04:32 haskell, lisp 2010-09-20T19:04:44 cYmen: ghc 6.8.2 is working already, and i think scheme is too 2010-09-20T19:04:46 esoteric in this case means awesome ;) 2010-09-20T19:04:55 dunno what flavor of scheme 2010-09-20T19:05:11 and i don't know about common lisp 2010-09-20T19:06:13 jmcarthur: isn't it Racket ? 2010-09-20T19:06:19 *** pc is now known as xvinyl 2010-09-20T19:06:29 cYmen: pretty much my workflow, once my bot is working at all, will be to test locally with the java engine (or one of these engines that have been posted to the forums), then test via tcp, then upload to the server for testing, where i am more critical of passing at each stage 2010-09-20T19:06:54 xvinyl: sounds right, but i don't know my schemes very well 2010-09-20T19:07:02 Hmm, these lazy admins haven't added new starter packages to site 2010-09-20T19:07:12 silly admins 2010-09-20T19:07:40 it's a shame though, many languages are supported and most of them already have startup packages 2010-09-20T19:08:06 everybody on the mailing list, i would like your support on the binaries issue (if you actually support my proposal, of course) 2010-09-20T19:08:15 I mean... most of hte job is already done on the admin's side 2010-09-20T19:09:49 jmcarthur: binary issue? 2010-09-20T19:10:42 cYmen: i'm trying to convince the maintainers/admins to allow arbitrary binaries for submission 2010-09-20T19:10:51 I vote 1! 2010-09-20T19:11:05 jmcarthur: How do you propose they do that in a safe manner? 2010-09-20T19:11:17 cYmen: how do you think they are running the code already? ;) 2010-09-20T19:11:28 Well, I would use a bunch of vms. :) 2010-09-20T19:11:34 cYmen: i would too 2010-09-20T19:11:50 cYmen: they are not currently doing that (bad bad bad!), but it makes no difference whether the source is available either way 2010-09-20T19:11:55 hmz 2010-09-20T19:12:14 how would you guys model threat ? i now choose a full out everything the enemy has got attack 2010-09-20T19:12:18 cYmen: it's just a chroot with some strict resource caps and iptables settings right now 2010-09-20T19:12:37 which makes it so risk averse that it can't flank 2010-09-20T19:12:40 bartwe: can you explain more clearly? 2010-09-20T19:12:46 mkay 2010-09-20T19:12:49 cYmen: well, and of course users with basically no privs 2010-09-20T19:13:29 sigh: when moving a fleet from A to B, i calculate the worstcase for both places. ie. the enemy sends everything, how much would my income be if i did, or did not send a force 2010-09-20T19:13:44 which makes it very cautious 2010-09-20T19:13:55 bartwe: ahh, i see... atm, my bot only cares about incoming fleets... this makes it very aggressive 2010-09-20T19:14:04 bartwe: that's very conservative indeed 2010-09-20T19:14:06 and leaves big weaknesses :P 2010-09-20T19:14:07 however, an all out attack would be surely overcome 2010-09-20T19:14:33 maybe i should include an all-out defense too 2010-09-20T19:14:57 what is an all-out defense? 2010-09-20T19:15:16 assume all my units move to the locations the enemy moves all his units to 2010-09-20T19:15:29 wouldn't that lock all your planets if you have less ships 2010-09-20T19:15:31 all ships move outside the playfield because they are scared 2010-09-20T19:15:50 bartwe: i would consider the minimum time it takes for the enemy to send a minimal force large enough to overtake my forces at that planet before i can minimally defend it, i think. i wouldn't worry about an all out rush 2010-09-20T19:15:57 sigh: if you have less ships, and less growth, and that happens, you lose anyways 2010-09-20T19:16:03 but then, i haven't experimented with a heuristic-only bot at all 2010-09-20T19:16:19 bartwe: plenty of games where a bot comes back from less ships and less growth 2010-09-20T19:16:19 where do you guys get your strategy ideas and such? 2010-09-20T19:16:20 ^^ accounting for growth i mean 2010-09-20T19:16:28 cYmen: from our heads! 2010-09-20T19:16:36 cYmen: thinking, watching other bots, and talking in the channel 2010-09-20T19:16:40 jmcarthur: that is what the .8 series does, but i can't break the top ranks anymore 2010-09-20T19:17:11 jmcarthur: yep, you need to planet trade when growth rates are advantageous 2010-09-20T19:17:46 jmcarthur: otherwise the heuristic you gave is still too cautious 2010-09-20T19:17:53 xvinyl: you're right 2010-09-20T19:18:13 i should actually keep my mouth shut. you guys have thought a lot more about heuristics than i have 2010-09-20T19:18:48 jmcarthur: i didn't write my bot yet :) 2010-09-20T19:19:12 nor i 2010-09-20T19:19:27 jmcarthur: my heuristics are suprisingly stupid 2010-09-20T19:19:46 but my bot will be a searchbot with only some specialized heuristics for it. my heuristics wouldn't be applicable for a standalone bot 2010-09-20T19:19:53 i'll have to learn to write correct Haskell first... 2010-09-20T19:20:33 hehe 2010-09-20T19:20:36 Learn You A Haskell For Great Bots 2010-09-20T19:20:44 lol yes 2010-09-20T19:20:57 it's hard to learn a language like haskell at the same time that you are trying to solve a difficult problem like this 2010-09-20T19:20:57 *** phreeza_ has quit IRC (Quit: Computer has gone to sleep) 2010-09-20T19:21:23 well, i've been trying to learn Haskell for a few months now 2010-09-20T19:21:40 just never had the opportunity to use it for a real project 2010-09-20T19:21:48 ah 2010-09-20T19:21:50 well good luck! 2010-09-20T19:21:56 actually it's a great way to learn haskell, but time to get your bot going will be rather large 2010-09-20T19:22:00 thanks :) 2010-09-20T19:22:19 dwchandler: sure. it's great for learning haskell, not for making a bot ;) 2010-09-20T19:22:29 :) 2010-09-20T19:22:36 well, we still have more than 2 months so... 2010-09-20T19:22:41 yeah 2010-09-20T19:22:56 meanwhile I can follow the latest strategies here :) 2010-09-20T19:23:05 hehe 2010-09-20T19:23:17 we should have bot explanation sessions :) 2010-09-20T19:23:28 everybody gets to present their bot to give the others ideas 2010-09-20T19:23:50 i shan't be revealing my strategies here! :P 2010-09-20T19:24:15 jmcarthur: is your bot on tcp or the main site? 2010-09-20T19:24:31 cYmen: on a somewhat related note, that's more or less what happened to the Netflix contest 2010-09-20T19:24:33 not done yet 2010-09-20T19:24:39 not *ready* yet i mean 2010-09-20T19:24:45 i should just add that to my nick 2010-09-20T19:24:47 xvinyl: What is the Netflix contest? 2010-09-20T19:24:52 bah, we can't even look at your games to steal your ideas :( 2010-09-20T19:24:56 jmcarthur_notready 2010-09-20T19:25:24 i wanted to do the netflix contest, but i never got a good chance 2010-09-20T19:25:29 not that i would have won anyway 2010-09-20T19:25:32 *** Naith has joined #aichallenge 2010-09-20T19:25:46 cYmen: people would submit algorithms to predict the score a user would give to a movie 2010-09-20T19:26:26 oh, neat 2010-09-20T19:26:30 Was it interesting? 2010-09-20T19:26:31 cYmen: it's mostly datamining from a big (user, movie) matrix they give you as a training set 2010-09-20T19:26:43 *** makg has quit IRC (Ping timeout: 252 seconds) 2010-09-20T19:26:54 cYmen: the goal was to one-up netflix's current algorithm by 10%, iirc. it took *years* 2010-09-20T19:27:22 cYmen: well I didn't enter the contest but the last 2 remaining teams were actually bunches of teams that had joined forces 2010-09-20T19:27:42 i was rooting for the psycologist working from his garage 2010-09-20T19:27:49 *** ath888 has quit IRC (Ping timeout: 252 seconds) 2010-09-20T19:27:50 jmcarthur: hehe 2010-09-20T19:28:28 haha, cool 2010-09-20T19:28:39 it was brilliant on netflix's part, too. they only had to spend a million on it not counting advertising and overhead 2010-09-20T19:28:40 cYmen: in the end they used mixtures of the results given by several algorithms 2010-09-20T19:28:53 jmcarthur: very true 2010-09-20T19:28:59 *** HakanD has joined #aichallenge 2010-09-20T19:29:12 they got tremendous advertisement/attention from this 2010-09-20T19:29:50 not to mention a 10% increase in the accuracy of their targetted inter-site "ads" 2010-09-20T19:29:59 *intra-site 2010-09-20T19:30:11 yep 2010-09-20T19:30:32 well, in the end i don't know if the winning algo was very scalable 2010-09-20T19:30:50 even if only as a batch job i think it was probably worth using 2010-09-20T19:31:04 10% is a big number in the grand scheme of things 2010-09-20T19:31:08 i think they mostly did this for the ideas their researchers could get off the submissions 2010-09-20T19:31:19 not necessarily the exact algorithms 2010-09-20T19:31:25 i wonder if they got any hires out of it 2010-09-20T19:31:28 if you see what i mean 2010-09-20T19:31:43 *** ath888 has joined #aichallenge 2010-09-20T19:31:45 yeah i see what you mean 2010-09-20T19:33:42 *** _Utkarsh <_Utkarsh!~admin@117.201.84.110> has joined #aichallenge 2010-09-20T19:33:49 *** HakanD has quit IRC (Ping timeout: 265 seconds) 2010-09-20T19:34:00 *** smellyhippy has quit IRC (Ping timeout: 252 seconds) 2010-09-20T19:34:22 *** Mathnerd314 has quit IRC (Disconnected by services) 2010-09-20T19:35:12 *** Mathnerd314_ has joined #aichallenge 2010-09-20T19:35:14 *** Mathnerd314_ is now known as Mathnerd314 2010-09-20T19:35:16 *** Utkarsh has quit IRC (Ping timeout: 265 seconds) 2010-09-20T19:35:42 *** delt0r_ has joined #aichallenge 2010-09-20T19:35:46 *** blkperl has quit IRC (Read error: Connection reset by peer) 2010-09-20T19:36:34 *** mega1 has quit IRC (Ping timeout: 240 seconds) 2010-09-20T19:36:34 *** delt0r___ has quit IRC (Read error: Connection reset by peer) 2010-09-20T19:36:35 *** ratatata has quit IRC (Quit: Page closed) 2010-09-20T19:36:40 *** mega1 has joined #aichallenge 2010-09-20T19:38:13 *** shades has quit IRC (Ping timeout: 276 seconds) 2010-09-20T19:38:15 Is the mailing list on gmane? 2010-09-20T19:38:22 *** rtaycher has quit IRC (Ping timeout: 276 seconds) 2010-09-20T19:38:23 *** shades has joined #aichallenge 2010-09-20T19:38:30 *** rtaycher has joined #aichallenge 2010-09-20T19:39:10 I can't even find the damn mailing list. :P 2010-09-20T19:39:13 no, it's a google groups list. it's also private for contributors only :( 2010-09-20T19:39:17 *** naglijo has joined #aichallenge 2010-09-20T19:39:27 or at least you have to be approved first 2010-09-20T19:39:41 i guess non-contributors are probably allowed. j3camero has to say so 2010-09-20T19:39:47 *** smellyhippy has joined #aichallenge 2010-09-20T19:39:47 *** smellyhippy has joined #aichallenge 2010-09-20T19:42:05 *** blkperl has joined #aichallenge 2010-09-20T19:43:09 *** delt0r___ has joined #aichallenge 2010-09-20T19:43:40 ok 2010-09-20T19:44:52 *** delt0r_ has quit IRC (Ping timeout: 276 seconds) 2010-09-20T19:47:41 *** jmreardon_ has joined #aichallenge 2010-09-20T19:49:36 *** jmreardon has quit IRC (Ping timeout: 245 seconds) 2010-09-20T19:49:36 *** jmreardon_ is now known as jmreardon 2010-09-20T19:56:25 Top 10 players: kcm1700(3509), McLeopold(3493), EBraun(3385), sighbot(3374), adumlah(3354), dmj111(3339), montanalow(3337), paulus(3323), Tracer(3316), petesimard(3275) 2010-09-20T19:56:26 snatching defeat from the hands of victory 2010-09-20T19:56:26 http://72.44.46.68/canvas?game_id=66397 2010-09-20T19:57:12 yay there is some actual discussion about binaries on the mailing list 2010-09-20T19:57:21 or at least jeff has responded apparently positively 2010-09-20T19:58:09 *** Naith has quit IRC (Read error: Connection reset by peer) 2010-09-20T19:59:24 *** Naith has joined #aichallenge 2010-09-20T20:00:54 *** mightybyte has quit IRC (Quit: leaving) 2010-09-20T20:01:02 *** justin_pdx has quit IRC (Quit: justin_pdx) 2010-09-20T20:01:24 *** mightybyte has joined #aichallenge 2010-09-20T20:01:40 *** justin_pdx has joined #aichallenge 2010-09-20T20:02:17 *** jmreardon has quit IRC (Quit: jmreardon) 2010-09-20T20:12:49 well yay 2010-09-20T20:14:54 hm...does google groups provide news server access? 2010-09-20T20:14:59 nntp or whatever that was again 2010-09-20T20:22:52 *** naglijo_ has joined #aichallenge 2010-09-20T20:23:23 *** ath888 has quit IRC (Quit: Page closed) 2010-09-20T20:24:04 You WIN against bartwe.10.11 2010-09-20T20:24:48 hmm, why does a bartwe bot has such a low rank? 2010-09-20T20:25:40 *** naglijo has quit IRC (Ping timeout: 276 seconds) 2010-09-20T20:25:40 *** n2kra has joined #aichallenge 2010-09-20T20:27:58 Naktibalda: new algorithm 2010-09-20T20:28:15 the cost function is either too pessimistic or too optimistic 2010-09-20T20:28:18 need fine tning 2010-09-20T20:44:20 *** Waco has joined #aichallenge 2010-09-20T20:49:36 Is MyBot case sensitive? 2010-09-20T20:54:51 yes, it is 2010-09-20T20:56:25 Top 10 players: McLeopold(3496), kcm1700(3479), sighbot(3388), EBraun(3364), petesimard(3358), dmj111(3357), adumlah(3355), montanalow(3349), Tracer(3341), paulus(3326) 2010-09-20T21:00:46 *** Macuyiko has quit IRC (Ping timeout: 264 seconds) 2010-09-20T21:03:57 *** Blkt has quit IRC (Remote host closed the connection) 2010-09-20T21:06:24 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-20T21:06:37 *** xvinyl has joined #aichallenge 2010-09-20T21:10:28 *** xvinyl has quit IRC (Remote host closed the connection) 2010-09-20T21:14:06 *** justin_pdx has joined #aichallenge 2010-09-20T21:27:35 *** Naith has quit IRC (Quit: Naith) 2010-09-20T21:48:44 *** FlagCapper has quit IRC (Quit: Page closed) 2010-09-20T21:56:08 anyone know what specific jvm version is used on the contest server? 2010-09-20T21:56:26 Top 10 players: McLeopold(3492), kcm1700(3479), dmj111(3412), sighbot(3389), EBraun(3362), petesimard(3354), adumlah(3353), montanalow(3348), Tracer(3337), paulus(3326) 2010-09-20T21:59:39 *** naglijo_ has quit IRC (Ping timeout: 252 seconds) 2010-09-20T22:16:01 *** Appleman1234 has quit IRC (Remote host closed the connection) 2010-09-20T22:47:59 *** McLeopold_ has joined #aichallenge 2010-09-20T22:56:15 *** McLeopold_ has quit IRC (Quit: Page closed) 2010-09-20T22:56:26 Top 10 players: McLeopold(3497), kcm1700(3485), dmj111(3411), sighbot(3393), adumlah(3365), petesimard(3360), montanalow(3357), EBraun(3353), Tracer(3344), paulus(3328) 2010-09-20T22:57:42 *** naglijo has joined #aichallenge 2010-09-20T23:06:52 *** smhanov has quit IRC (Ping timeout: 252 seconds) 2010-09-20T23:07:59 *** Zeiris has joined #aichallenge 2010-09-20T23:10:13 *** smhanov has joined #aichallenge 2010-09-20T23:13:30 quiet here tonight 2010-09-20T23:19:24 whats the topic? 2010-09-20T23:19:55 potatos 2010-09-20T23:20:16 good old spuds 2010-09-20T23:23:27 anyone play the new tom clancy game 2010-09-20T23:23:30 ? 2010-09-20T23:24:00 kinda looks... well insane! 2010-09-20T23:30:59 anyone play that new game toejam and earl! looks awsome 2010-09-20T23:32:28 ... 2010-09-20T23:34:35 No Haskellers on the top 100. :( 2010-09-20T23:34:53 not *yet* ;) 2010-09-20T23:34:59 oh, there used to be 2010-09-20T23:35:03 i guess they dropped 2010-09-20T23:36:09 jmcarthur: Did you get time to submit your bot? 2010-09-20T23:38:44 whats everyone's rank? 2010-09-20T23:38:58 *** naglijo_ has joined #aichallenge 2010-09-20T23:38:58 on the chat 2010-09-20T23:39:02 I'm stuck somewhere in the 400s. :( 2010-09-20T23:39:07 on the chat? 2010-09-20T23:39:27 yea room 2010-09-20T23:39:33 ahhh 2010-09-20T23:40:11 *** _Utkarsh <_Utkarsh!~admin@117.201.84.110> has quit IRC (Ping timeout: 276 seconds) 2010-09-20T23:40:32 My poor bot's in the 1000's somewhere 2010-09-20T23:40:52 feel bad for it but it win's 2010-09-20T23:40:54 can you beat ragebot? 2010-09-20T23:40:59 *** naglijo has quit IRC (Ping timeout: 264 seconds) 2010-09-20T23:41:30 right now I'm working on trying to make it a real AI 2010-09-20T23:41:51 something that can actually know what to do 2010-09-20T23:42:00 road blocked 2010-09-20T23:42:03 bss03-AFK: no not yet 2010-09-20T23:42:21 Snowplan: yeah, you got to try a bunch of stuff and see what works out 2010-09-20T23:42:30 my idea is complex 2010-09-20T23:42:49 work in small steps, for example you might implement some simple defense if you haven't yet 2010-09-20T23:43:00 the example bots have no defense in them 2010-09-20T23:43:51 right, I mainly am working on strategy but my biggest flaw is sucking at programing. 2010-09-20T23:44:15 ah well, this is a good opportunity to practice and get better :) 2010-09-20T23:44:16 I'm using java 2010-09-20T23:44:27 exactly 2010-09-20T23:45:49 my sorting process sucks and it's ashame 2010-09-20T23:45:59 sorting process? 2010-09-20T23:46:24 *** Utkarsh has joined #aichallenge 2010-09-20T23:46:42 *** bss03-AFK is now known as bss03 2010-09-20T23:46:57 yea I am not sorting my planets correct 2010-09-20T23:47:08 nor other planets 2010-09-20T23:47:10 I think I finally have a bot that is better than my last submission.. finally. 2010-09-20T23:47:25 I kept making things more and more complex, and not really getting better -- or getting worse. 2010-09-20T23:47:30 Snowplan: you mean to determine which planets are valuable? 2010-09-20T23:47:52 right 2010-09-20T23:47:53 Today I simpled a few things up, and it seems to be working out. 2010-09-20T23:48:36 or which are vulnerable 2010-09-20T23:49:18 Snowplan: yup, that is difficult, even the top bots have trouble with it 2010-09-20T23:49:26 you just have to improve it bit by bit 2010-09-20T23:49:41 also distance 2010-09-20T23:49:43 try different things 2010-09-20T23:49:51 thanks for the encouragement 2010-09-20T23:49:55 :) 2010-09-20T23:51:03 *** Naith has joined #aichallenge 2010-09-20T23:51:28 bss03: yeah, keeping things simple is always a bonus, having a lot of special cases can be an indication you are missing the bigger picture 2010-09-20T23:53:21 Any tips on the elo logic of it? 2010-09-20T23:53:32 elo logic? 2010-09-20T23:53:46 It wasn't that there was a lot of special cases, it was just trying to be "too" smart and missing the point. 2010-09-20T23:53:57 My algorithms were too complex for what I was really trying to do. 2010-09-20T23:54:23 bss03: yeah, I've come across that too :) 2010-09-20T23:56:01 Snowplan: I assign a value to a planet based on it's growthRate * (# of turns) - shipsToConquer 2010-09-20T23:56:07 i hope that doesn't happen to me 2010-09-20T23:56:16 trying to be too smart, that is 2010-09-20T23:56:27 Top 10 players: McLeopold(3511), kcm1700(3504), dmj111(3433), sighbot(3372), EBraun(3371), adumlah(3367), petesimard(3362), montanalow(3354), paulus(3334), Tracer(3288) 2010-09-20T23:56:45 Snowplan: I don't include distance directly in the calculation, but "# of turns" is decreased by the time it would take me to get shipsToConquer there. 2010-09-20T23:56:58 *** pjreddie has quit IRC (Ping timeout: 252 seconds) 2010-09-20T23:57:32 Unfortunately, when I'm facing a good bot, shipsToConquer is always a gross underestimate the way I'm calculating it now. :( 2010-09-20T23:58:05 Is it possible to check the PlayGame source? 2010-09-20T23:58:09 Snowplan: also you might want to read this: http://ai-contest.com/forum/viewtopic.php?f=17&t=593 2010-09-20T23:58:17 it has a lot of good tips 2010-09-20T23:58:57 Zarathust: You might check the Google Code project and see if it has the PlayGame source. 2010-09-20T23:59:37 Zarathust: PlayGame.jar isn't used for it official engine though, IIRC. ISTR the actuall engine is written in python. 2010-09-20T23:59:58 where can I get it?