2010-09-27T00:00:23 *** delt0r___ has joined #aichallenge 2010-09-27T00:01:21 jmcarthur: True... I have been pushing FsCheck every chance I get though :) 2010-09-27T00:01:30 *** delt0r_ has quit IRC (Ping timeout: 240 seconds) 2010-09-27T00:02:42 * dmj111 lights out 2010-09-27T00:02:54 *** Olathe has quit IRC (Ping timeout: 272 seconds) 2010-09-27T00:04:34 *** davidjliu has quit IRC (Quit: Page closed) 2010-09-27T00:22:04 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-09-27T00:25:29 *** iFire has joined #aichallenge 2010-09-27T00:27:55 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-09-27T00:38:07 *** JCS^ has quit IRC (Quit: \x00) 2010-09-27T00:43:00 *** hornairs has quit IRC (Remote host closed the connection) 2010-09-27T00:43:55 *** troutwine has quit IRC (Ping timeout: 245 seconds) 2010-09-27T00:49:14 my ranking is taking forever to go up - http://ai-contest.com/profile.php?user_id=3911 2010-09-27T00:53:02 *** rwa has quit IRC (Quit: rwa) 2010-09-27T00:57:38 Top 10 players: dmj111(3619), adumlah(3508), bix0r4ever(3508), sequoh(3498), davidjliu(3451), McLeopold(3440), kcm1700(3417), vinx33(3409), Rexxar(3395), sighbot(3394) 2010-09-27T00:58:28 I now have 9 wins with 1267 ranking - last guy was 860 2010-09-27T00:59:11 you were sort of unfortunate that your first 3 were all under 2000 2010-09-27T00:59:25 they're going to drag you down for a bit 2010-09-27T00:59:34 you mean my first 3 were over 2000? 2010-09-27T00:59:56 well whichever way you're sorting it down from 1 or up :) 2010-09-27T01:00:20 doing good, seen 1 bug though and want to see it lose so I can fix it 2010-09-27T01:00:37 99/100 dualbot and 96/100 ragebot 2010-09-27T01:00:39 *** keyne has joined #aichallenge 2010-09-27T01:00:49 :) 2010-09-27T01:00:51 matching algorithm seems a bit screwed... I'm at 10th and in the 3 games I've played people ranked 981 and 876 2010-09-27T01:01:00 Cyndre: not bad 2010-09-27T01:01:15 to get the last few against ragebot, you need pretty specific code 2010-09-27T01:01:34 it just ranked me against a 1600 2010-09-27T01:01:35 not terribly urgent until you have other stuff sorted out, imo 2010-09-27T01:01:40 sigh: at least the algorithm checked into svn for the distributed servers is horrendous 2010-09-27T01:01:40 http://ai-contest.com/visualizer.php?game_id=4816199 2010-09-27T01:02:16 Janzert: haven't looked at the code, what do they do? 2010-09-27T01:02:17 I'm hoping danielvf has actually already changed that but just not committed the change yet 2010-09-27T01:02:24 wow 2010-09-27T01:02:24 http://ai-contest.com/visualizer.php?game_id=4816199 2010-09-27T01:02:30 sorry 2010-09-27T01:02:48 choose first player randomly then choose within +-20 *submissions* 2010-09-27T01:02:58 there can be planets 1 apart 2010-09-27T01:03:18 what does "+-20 *submissions*" mean? 2010-09-27T01:03:54 choose the second player randomly from the 20 submissions before or after the first player chosen 2010-09-27T01:04:14 Zannick: yeah... I've heard several times from admins that the python script guards against it, but I looked at the code and I can't see it 2010-09-27T01:04:18 Janzert: the ranked submission, you mean 2010-09-27T01:04:27 sigh: i examined it 2010-09-27T01:04:28 it does have a guard against planets that are too close, but I don't see how it stops 1 distance planets 2010-09-27T01:04:34 it doesn't 2010-09-27T01:04:45 cool, I'm not crazy :) 2010-09-27T01:04:58 no, just submission order from the way I'm reading it 2010-09-27T01:05:08 hopefully I'm wrong 2010-09-27T01:05:13 the too close guard goes off for 2 1-growth planets if they are closer than .8 2010-09-27T01:05:15 http://code.google.com/p/ai-contest/source/browse/trunk/planet_wars/www/api_get_matchup.php?spec=svn376&r=376 2010-09-27T01:05:27 so, .85 distance is fine, distance 1 2010-09-27T01:05:45 furthermore, the guard isn't applied against the first three planets 2010-09-27T01:05:58 Zannick: agree 2010-09-27T01:06:09 I have a feeling resubmiting my bot until I draw my first matchup with someone in the top 100 is going to be faster than waiting to play some games. 2010-09-27T01:06:20 so there is a .5454% chance that the starting planets are each 1 away from the center planet 2010-09-27T01:06:38 and a quarter of that possibility that the starting planets are 1 apart from each other! 2010-09-27T01:07:48 Janzert: the submissions array is ordered by rank 2010-09-27T01:08:22 oh god - I fought a guy lower then me and went up in rank..... 2010-09-27T01:08:36 yeah, active submissions, ordered by rank 2010-09-27T01:08:42 ahh, ok. I see now. I missed the last line of the query 2010-09-27T01:08:46 :) 2010-09-27T01:08:58 *** troutwine has joined #aichallenge 2010-09-27T01:09:10 it's actually the inner select, sort within userid, max submission id from those 2010-09-27T01:09:15 but anyway 2010-09-27T01:09:32 not the "order by rank desc"? 2010-09-27T01:10:22 wait 2010-09-27T01:10:34 i'm looking at tournament_manager.py 2010-09-27T01:10:36 actually the only inner select I see is the max(leaderboard_id) 2010-09-27T01:10:57 ahh, yeah that is quite a bit different 2010-09-27T01:11:41 oookay 2010-09-27T01:11:53 i no longer know how he's picking matches 2010-09-27T01:12:40 anyway, i think i shall submit a fix for the map generator 2010-09-27T01:13:17 should have an undefeated bot link on site that shows every bot with 0 loss's 2010-09-27T01:14:22 except it will encourage people to resubmit as soon as they get a loss 2010-09-27T01:14:53 wow, I was 24 on tcp 2010-09-27T01:17:10 undefeated doesn't mean anything unless you are #1 2010-09-27T01:17:29 we already have a list for that :P 2010-09-27T01:19:50 * Janzert just noticed the rankings have broken past the 3000 elo point spread 2010-09-27T01:21:10 what is the significance of that? 2010-09-27T01:22:04 nothing in and off itself, just the wider the spread the larger skill differential from top to bottom 2010-09-27T01:22:39 i see 2010-09-27T01:23:02 3000 means 6 steps where each step should beat the next 95% of the time 2010-09-27T01:23:33 get to 3500 and it's 7 :) 2010-09-27T01:24:32 cool :) 2010-09-27T01:25:55 so bots should beat opponents 500 points lower 95% of the time? 2010-09-27T01:26:00 hmm, it'd be interesting to see the elo spread from top rank to median rank over time 2010-09-27T01:26:21 in general yes, although any specific case will vary of course 2010-09-27T01:26:48 http://www.3dkingdoms.com/chess/elo.htm can give you the elo difference for any specific winning percentage 2010-09-27T01:26:52 coo, I don't know much about elo... now I can make more sense of the ranks :) 2010-09-27T01:28:07 I don't think the rating change calculator part of that page will match well with what happens in the contest though since bayeselo is used and is quite a bit more complex 2010-09-27T01:30:24 btw Janzert, what is your username in the tournament? 2010-09-27T01:30:52 Janzert, I just haven't made a submission yet :P 2010-09-27T01:31:12 ah :) 2010-09-27T01:32:02 I did actually work on writing a bit of code for it last night though :) 2010-09-27T01:32:49 and see up and comming bots 2010-09-27T01:33:09 Janzert - of course make sure they cant duplicate upload 2010-09-27T01:33:10 Janzert: the competition gets harder the longer you wait 2010-09-27T01:33:50 sigh: yes, it's starting to give me confidence that this game isn't to shallow 2010-09-27T01:34:03 another reason I like seeing the elo spread increase 2010-09-27T01:34:18 Janzert: you thought otherwise? 2010-09-27T01:34:44 tron was easy to see that it reduced to a NP hard problem, this one isn't obviously that way 2010-09-27T01:34:57 yea 2010-09-27T01:35:06 it could be that a relatively simple strategy dominated all others 2010-09-27T01:35:13 Im upto 864 undefeated :) 2010-09-27T01:35:26 nice 2010-09-27T01:36:42 Javache: yeah, this game is very intricate... I don't think anyone has a clear idea of what is good yet 2010-09-27T01:37:07 well, maybe the bots right at the top... I can't speak for them 2010-09-27T01:37:10 http://ai-contest.com/visualizer.php?game_id=4816744 <--- thats whats good about it 2010-09-27T01:37:12 :) 2010-09-27T01:37:58 yep, I was worried at first that after getting past the surface complexity of, basically, the high branching factor there might not be much depth 2010-09-27T01:37:59 Cyndre: you are just bullying people :P 2010-09-27T01:38:34 :) 2010-09-27T01:38:52 http://ai-contest.com/visualizer.php?game_id=4815022 watch turn 18 to 31 for the sucker punch 2010-09-27T01:40:49 heh, looks a little dangerous though it worked quite nicely 2010-09-27T01:41:02 how is it dangerous? 2010-09-27T01:41:28 I send more cause clearly Im just not checking right - but get a planet for 3 ships, and the enemy loses 50 2010-09-27T01:41:54 all those ships completely committed from all the way across the map, no way to change your mind 2010-09-27T01:42:07 3 ships if done right 2010-09-27T01:42:13 but remember this is someone who has yet to write his first submission talking :) 2010-09-27T01:42:37 nah, I agree with Janzert, but it would take a smarter bot than any currently out there to take advantage 2010-09-27T01:42:43 but you see the one Im talking about - the center planet? 2010-09-27T01:43:14 *** ghostsoft has quit IRC (Quit: ghostsoft) 2010-09-27T01:43:15 ahh, yes that is nicely done 2010-09-27T01:43:30 thats where I won that game 2010-09-27T01:50:28 *** rwa has joined #aichallenge 2010-09-27T01:53:21 *** mega1 has joined #aichallenge 2010-09-27T01:53:44 *** Utkarsh has quit IRC (Read error: Connection reset by peer) 2010-09-27T01:57:38 Top 10 players: dmj111(3620), adumlah(3522), bix0r4ever(3506), sequoh(3472), davidjliu(3443), McLeopold(3435), kcm1700(3425), vinx33(3402), Rexxar(3402), sighbot(3390) 2010-09-27T02:02:10 *** Utkarsh has joined #aichallenge 2010-09-27T02:05:05 okay, i think i'm done looking through the issues list for things to close/comment on 2010-09-27T02:05:11 there are so many issues D: 2010-09-27T02:05:45 should add that as an issue 2010-09-27T02:06:15 assignee: cameron.jp summary: clean up issues list 2010-09-27T02:06:25 *** Utkarsh has quit IRC (Ping timeout: 245 seconds) 2010-09-27T02:06:30 :) 2010-09-27T02:11:16 *** Itkovian has joined #aichallenge 2010-09-27T02:11:17 *** Utkarsh has joined #aichallenge 2010-09-27T02:12:04 no game in like 30 mins 2010-09-27T02:14:16 * Janzert 's offer to help with the issue list is still in play, should j3camero ever come back :) 2010-09-27T02:14:34 Cyndre: my last game was after a wait of over 3 hours 2010-09-27T02:15:07 Ive had 13.... but onyl rank 828 and dont like replacing before I lose 2010-09-27T02:15:17 which I just got 2010-09-27T02:15:23 heh 2010-09-27T02:15:29 http://ai-contest.com/visualizer.php?game_id=4818779 2010-09-27T02:16:04 at least it's a loss to a guy that's been hanging out in the top 10 2010-09-27T02:16:22 what? 2010-09-27T02:16:43 I just resubmitted 2010-09-27T02:16:53 http://ai-contest.com/visualizer.php?game_id=4818825 2010-09-27T02:17:03 from a couple of hours ago: contestbot: Top 10 players: adumlah(3475), bix0r4ever(3471), dmj111(3455), sequoh(3454), McLeopold(3430), davidjliu(3384), felixcoto(3380), sighbot(3358), kcm1700(3322), vinx33(3311) 2010-09-27T02:17:41 damn, I scrapped that version cause he lost 2010-09-27T02:17:47 hahaha 2010-09-27T02:18:22 my new bot is better and fix's a bug where it leaves a planet with 0 enemy ships - so kind of happy to have lost 2010-09-27T02:18:46 :) 2010-09-27T02:18:55 it should have been obvious from the game that felixcoto was not ranked under 1000 :P 2010-09-27T02:19:19 MyBot vs DualBot 99/100 Avg Turns: 57.0, MyBot vs RageBot 96/100 Avg Turns: 48.0 2010-09-27T02:20:02 Cyndre: leaves a planet with 0 ships? 2010-09-27T02:20:27 I was cleaning my code and removing all the many + 1 or - 1 I have put in to solve little bugs 2010-09-27T02:20:34 and took out 1 that had to be there :) 2010-09-27T02:20:37 is there an invalid move on the server that tcp doesn't catch? 2010-09-27T02:20:43 I keep losing for unexplained reasons on the main server 2010-09-27T02:20:50 I've cut the time down to tiny amounts 2010-09-27T02:20:55 printing out something with 3 spaces 2010-09-27T02:21:11 rwa: parse the playback feed it into your bot and look at the output 2010-09-27T02:21:11 start your logging like print "s o l: logMessageHere" 2010-09-27T02:21:44 http://72.44.46.68/getplayer?player=Cyndre thaty bug fix made a big diff 2010-09-27T02:22:15 sigh: how do I get the playback data 2010-09-27T02:22:43 rwa: http://www.ai-contest.com/forum/viewtopic.php?f=17&t=628 2010-09-27T02:23:04 contestbot should have a command for that, this is like the 10th time I've posted this link :P 2010-09-27T02:23:05 sigh: Run as fast as you can and don't look back. 2010-09-27T02:23:16 already am, contestbot 2010-09-27T02:23:28 ty 2010-09-27T02:24:56 *** zl has joined #aichallenge 2010-09-27T02:26:16 contestbot doesn't understand that this isn't a command 2010-09-27T02:26:17 Zannick: An error occured while trying to show the previous error. 2010-09-27T02:27:13 !yay contestbot! 2010-09-27T02:27:13 Double rainbow all the way! 2010-09-27T02:28:40 *** Itkovian has quit IRC (Quit: Itkovian) 2010-09-27T02:35:21 *** wvdschel has joined #aichallenge 2010-09-27T02:41:15 *** otherAntimatroid has joined #aichallenge 2010-09-27T02:55:54 *** kjeldahl has joined #aichallenge 2010-09-27T02:56:27 *** Utkarsh has quit IRC (Ping timeout: 272 seconds) 2010-09-27T02:57:39 Top 10 players: dmj111(3611), adumlah(3516), bix0r4ever(3504), Rexxar(3465), sequoh(3465), davidjliu(3439), kcm1700(3418), McLeopold(3412), sighbot(3384), vinx33(3380) 2010-09-27T02:58:55 *** Arthur has joined #aichallenge 2010-09-27T03:00:14 *** Appleman1234 has joined #aichallenge 2010-09-27T03:01:14 *** keyne has quit IRC (Read error: Connection reset by peer) 2010-09-27T03:01:15 *** Utkarsh has joined #aichallenge 2010-09-27T03:05:14 my new bot is up - MyBot vs DualBot 100/100 Avg Turns: 58.0, MyBot vs RageBot 96/100 Avg Turns: 50.0 2010-09-27T03:06:58 before you were beating them quicker, what happened? :P 2010-09-27T03:07:28 not sure 2010-09-27T03:07:58 I had a place with this comment #we dont have enough - send them all and continue planning 2010-09-27T03:08:02 lmao 2010-09-27T03:08:13 so I limited to growthRate of target planet for now 2010-09-27T03:08:30 hoping it will fix my stupid loss's 2010-09-27T03:08:32 heh 2010-09-27T03:08:37 *** troutwine has quit IRC (Ping timeout: 240 seconds) 2010-09-27T03:08:57 my bot is much more conservative, it only sends if it naively thinks it can win the planet 2010-09-27T03:09:12 sigh: it's hard to get that right aye 2010-09-27T03:09:35 antimatroid: yes, winning is hard :P 2010-09-27T03:09:35 i think after my dismal duct taping efforts yesterday i'm giving up on throwing new bots together every few days 2010-09-27T03:09:47 just beat the number 2 bot on the tcp server 2010-09-27T03:09:52 nice 2010-09-27T03:09:53 antimatroid: what do you mean? 2010-09-27T03:09:57 http://72.44.46.68/getplayer?player=Cyndre 2010-09-27T03:09:59 :) 2010-09-27T03:10:22 sigh: i'm spending time just thinking through the problem again, and implementing longer ideas locally and testing those thoroughly etc. 2010-09-27T03:10:45 yeah, fair enough 2010-09-27T03:10:51 plus i have more other work to do than i can poke a stick at 2010-09-27T03:10:55 !rainbow I found a new command! 2010-09-27T03:10:55 sigh: I found a new command! 2010-09-27T03:11:27 !rainbow happy colours! 2010-09-27T03:11:27 antimatroid: happy colours! 2010-09-27T03:13:17 *** jaspervdj has joined #aichallenge 2010-09-27T03:15:50 *** iFire has joined #aichallenge 2010-09-27T03:16:38 *** Naith has joined #aichallenge 2010-09-27T03:23:32 *** Utkarsh has quit IRC (Ping timeout: 252 seconds) 2010-09-27T03:27:05 *** troutwine has joined #aichallenge 2010-09-27T03:28:51 *** aerique has joined #aichallenge 2010-09-27T03:30:05 *** Utkarsh has joined #aichallenge 2010-09-27T03:40:39 *** sigh has quit IRC (Remote host closed the connection) 2010-09-27T03:46:02 !rainbow I'm colour blind. What does this do exactly? 2010-09-27T03:46:02 shades: I'm colour blind. What does this do exactly? 2010-09-27T03:47:15 I don't think you understand how colour blindness works 2010-09-27T03:47:45 *** SophieBC has joined #aichallenge 2010-09-27T03:47:47 *** wvdschel has quit IRC (Quit: Leaving) 2010-09-27T03:48:13 I have the jist of it 2010-09-27T03:48:29 red/green green/blue or all three or something 2010-09-27T03:52:04 *** Naith has quit IRC (Read error: Connection reset by peer) 2010-09-27T03:52:24 *** Naith has joined #aichallenge 2010-09-27T03:54:11 *** troutwine has quit IRC (Ping timeout: 240 seconds) 2010-09-27T03:54:20 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-09-27T03:55:50 *** Appleman1234 has quit IRC (Ping timeout: 240 seconds) 2010-09-27T03:56:20 *** Appleman1234 has joined #aichallenge 2010-09-27T03:57:39 Top 10 players: dmj111(3617), adumlah(3523), bix0r4ever(3512), Rexxar(3475), sequoh(3473), davidjliu(3454), kcm1700(3448), McLeopold(3418), sighbot(3391), vinx33(3374) 2010-09-27T04:00:12 *** bhasker has joined #aichallenge 2010-09-27T04:00:26 hi 2010-09-27T04:02:13 hi 2010-09-27T04:02:45 http://72.44.46.68/canvas?game_id=184108 woot 2010-09-27T04:03:00 nice game 2010-09-27T04:03:31 *** Appleman1234 has quit IRC (Ping timeout: 240 seconds) 2010-09-27T04:03:42 my bot is finally doing somewhat decently 2010-09-27T04:03:52 nice, rank? 2010-09-27T04:03:57 *** Appleman1234 has joined #aichallenge 2010-09-27T04:04:02 on the unofficial server at 73 2010-09-27T04:04:09 haven't uploaded it to the official server yet 2010-09-27T04:04:26 i am using python 2.6 and the official one is 2.5 so i have to make some changes before it will work on the official 2010-09-27T04:04:41 http://72.44.46.68/getplayer?player=bhaskerv20.1 2010-09-27T04:04:44 damn, ask them to upgrade 2010-09-27T04:05:26 nah should be a minor change, 2.5 doesn't have the PriorityQueue class so i just need to reimplement it using a heapq in 2.5 2010-09-27T04:05:39 or just change my code to use heapq instead of PriorityQueue 2010-09-27T04:05:52 i think thats the only thing in my code thats 2.6 specific 2010-09-27T04:06:38 pretty interesting to watch the top bots 2010-09-27T04:08:06 how is your bot doing Cyndre 2010-09-27T04:08:12 not to bad 2010-09-27T04:08:36 http://ai-contest.com/profile.php?user_id=3911 2010-09-27T04:08:49 nice 2010-09-27T04:08:49 http://72.44.46.68/getplayer?player=Cyndre 2010-09-27T04:09:27 thanks 2010-09-27T04:09:44 now that I have a competing bot again I can look at where Im losing the games and fix it :) 2010-09-27T04:09:52 most of the time I was ahead then fell behind :( 2010-09-27T04:10:02 but I am off to sleep - work in 6 hours, night 2010-09-27T04:10:49 night 2010-09-27T04:11:00 you're doing better than i, i managed to climb into bed just after 5am last night 2010-09-27T04:11:36 antimatroid your bot was ripping it up on the unofficial server a few days back 2010-09-27T04:11:50 *** di has joined #aichallenge 2010-09-27T04:11:55 yeah, i've added new things and made it way worse 2010-09-27T04:12:09 heh happens all the time 2010-09-27T04:12:13 hopefully with some refinement and a few other things added in the make it smarter and it'll be back near the top 2010-09-27T04:12:31 adum's bot is pretty interesting to watch 2010-09-27T04:12:42 yep 2010-09-27T04:12:53 but i have to put it on hold for the moment, i have an analysis test wednesday 2010-09-27T04:12:58 ah 2010-09-27T04:13:49 i have no idea how i'm going to get all my work done in the next two months 2010-09-27T04:13:57 i have exams finishing about 2 weeks before the end of the comp 2010-09-27T04:14:09 ouch 2010-09-27T04:14:25 and i took on 6 units before knowing the comp was coming up :( 2010-09-27T04:14:38 I beat him down tonight :) 2010-09-27T04:14:46 well i am just doing this for fun, since i work full time 2010-09-27T04:14:49 although one is unofficially, so i don't have to sit the exam right away 2010-09-27T04:15:08 http://ai-contest.com/visualizer.php?game_id=4822978 nice, just took out a rank 62 bot 2010-09-27T04:15:36 i'm a huge game theory fan and quite enjoy playing around with algorithsm/maths, so this kind of thing is very appealing to me :P 2010-09-27T04:15:39 *** Utkarsh has quit IRC (Ping timeout: 265 seconds) 2010-09-27T04:16:05 ah k 2010-09-27T04:16:11 my bot has been pretty dumb right now 2010-09-27T04:16:19 i am not doing any prediction of enemies moves 2010-09-27T04:16:23 its purely heuristics 2010-09-27T04:16:38 so no minmax/pruning etc 2010-09-27T04:16:46 anything you've seen of mine on official servers has been so far 2010-09-27T04:16:53 but i'm working on other methods on paper 2010-09-27T04:17:03 i tried a dp approach on the weekend, but was WAY too slow 2010-09-27T04:17:34 enlighten me what does DP mean:D 2010-09-27T04:17:39 dynamic programming 2010-09-27T04:17:41 I beat dualbot 100/100 58 turns avg, and ragebot 96/100 in 50 turns avg 2010-09-27T04:17:50 ah k 2010-09-27T04:17:56 Cyndre: try dual bot on map 44 with you as player 2 2010-09-27T04:17:57 *** hellman has joined #aichallenge 2010-09-27T04:18:29 they move to an actractive planet for you then 2010-09-27T04:18:31 *** amstan has quit IRC (Ping timeout: 245 seconds) 2010-09-27T04:18:35 and we had to fix up our bot getting owned on that 2010-09-27T04:18:55 461 2010-09-27T04:18:57 I win 2010-09-27T04:19:51 23 turns if the spots are swapped 2010-09-27T04:19:52 wow 2010-09-27T04:20:00 yep 2010-09-27T04:20:07 either you cane them or they cane you in that instance i find 2010-09-27T04:20:10 due to their first move 2010-09-27T04:20:51 what language do you use? 2010-09-27T04:21:04 who are you asking? I'm using c++ 2010-09-27T04:21:06 http://ai-contest.com/visualizer.php?game_id=4823162 2010-09-27T04:21:22 *** Utkarsh has joined #aichallenge 2010-09-27T04:23:06 my bot slaughters it self on that map 2010-09-27T04:23:21 my bot too.. 2010-09-27T04:23:24 :D 2010-09-27T04:23:38 39 turns 2010-09-27T04:23:46 er 28 turns, seat two kicks ass 2010-09-27T04:27:00 map 1 is biased 2010-09-27T04:27:25 *** Palmik has joined #aichallenge 2010-09-27T04:27:29 map 2 2010-09-27T04:27:47 on neutral planet (0 fleets) fly 2 fleets from other peoples at one time, whose planet will be in this case ? 2010-09-27T04:27:48 map 34 2010-09-27T04:30:25 any suggestions? 2010-09-27T04:31:09 *** Naktibalda has joined #aichallenge 2010-09-27T04:33:45 *** shades has quit IRC (Ping timeout: 264 seconds) 2010-09-27T04:34:02 hmmm, should the servers be running bots against each other from both starting positions? 2010-09-27T04:34:11 seems like it might be luck sometimes 2010-09-27T04:34:29 aren't both positions even? 2010-09-27T04:34:34 you would think 2010-09-27T04:34:50 but seems it doesn't always work that way based on ordering of planet numbers 2010-09-27T04:34:56 *** Itkovian has joined #aichallenge 2010-09-27T04:35:20 pherhaps the players could both "think" they are player 1 and then the engine switches one players moves? 2010-09-27T04:35:22 have you found a non-symetric map? 2010-09-27T04:35:27 read above 2010-09-27T04:35:39 I just came in 2010-09-27T04:35:47 *** jens has joined #aichallenge 2010-09-27T04:35:49 it's not about the map being asymetric, rather that the numbering of planets causes some to be picked before others 2010-09-27T04:36:04 i assume that is what is causing some variation in results where one would expect them to be identical 2010-09-27T04:36:05 *** jens is now known as JensT 2010-09-27T04:36:09 *** sigh has joined #aichallenge 2010-09-27T04:37:05 i actually like the idea of having the engine have both players think they are player 1 2010-09-27T04:37:22 then for the information they receive from each player, reverse it as though it was player 2, then send that to the other bot 2010-09-27T04:37:29 that way it is necessarily symmetric yes? 2010-09-27T04:37:50 damn vortex.* variants keep owning my bot:( 2010-09-27T04:38:58 anyone agree or disagree with that? 2010-09-27T04:40:17 I think we should be given a random numbered idea each round 2010-09-27T04:40:23 preferably irrational 2010-09-27T04:40:29 *id 2010-09-27T04:40:37 say what? 2010-09-27T04:40:58 I just came in, what's happening? 2010-09-27T04:41:05 i'm talking about the way in which results of reversing players doesn't always resiult in the same player winning from the same moves 2010-09-27T04:41:22 ignoring randomness, i think my way would fix that 2010-09-27T04:41:43 what is your way? 2010-09-27T04:42:13 *** JensT has quit IRC (Quit: Leaving.) 2010-09-27T04:42:31 have the players both think they're player 1, then engine swaps those moves to equivalent moves of player 2 and sends that information to the other bot, then engine arbitrarily has one player as 2 for it's own state information 2010-09-27T04:42:55 so each player thinks they're in the position of player 1, with symmetric maps that should work 2010-09-27T04:43:19 each player thinks they are player 1 right now... 2010-09-27T04:44:18 then how do we get variation? 2010-09-27T04:44:25 is the ceil integer function causing asymetries? 2010-09-27T04:44:29 i wouldn't have thought so 2010-09-27T04:44:46 asymettry is from the planet numbering 2010-09-27T04:44:50 cound you give an example of asymetry? 2010-09-27T04:44:56 but my way they aren't different 2010-09-27T04:45:20 Naktibalda: my bot performs differently against dual bot on map 44 dependent on which player it plays as 2010-09-27T04:45:31 wins both these days, but it's a different game 2010-09-27T04:45:55 dhartmei was saying the other day that he's seen such things a lot on tcp data 2010-09-27T04:45:59 dualbot doesn't care about distance so it picks equal planets based on id 2010-09-27T04:46:35 half the time it will make the stupid choice 2010-09-27T04:46:52 exactly, so the planet id's should be reversed if it's player 2 2010-09-27T04:47:00 such that it makes the "same" move 2010-09-27T04:47:06 oh, a special fix for dual bot :) 2010-09-27T04:47:06 that's my point, to account for planet numbering 2010-09-27T04:47:17 no, not for dual bot, that's an example 2010-09-27T04:47:38 i see what your saying, I'm just not sure why it matters 2010-09-27T04:48:01 because otherwise you might have 2 bots currently that would swap win/lose dependent on starting positions 2010-09-27T04:48:13 i would say its then reasonable that they should be played from both starting positions 2010-09-27T04:48:22 or use my method which forces symmetry in an extra way 2010-09-27T04:48:24 *** shades has joined #aichallenge 2010-09-27T04:48:24 they are the same starting positions 2010-09-27T04:48:37 yes, but the planet numbering isn't symmetric 2010-09-27T04:48:40 the planet ids are arbitrary and any reliance on them is the fault of the bot 2010-09-27T04:49:08 yes, and i don't intend for mine to depend on it, but wouldn't it be nicer to eliminate such possible discrepencies anyway? 2010-09-27T04:49:18 if the bot really wants, it can reorder the planet ids on internally 2010-09-27T04:49:40 ignore what it means for bots, think about nice consistent correct results 2010-09-27T04:49:51 it duplicates load of servers just to make dulabot behave more consistently :) 2010-09-27T04:50:03 it's not for dual bot 2010-09-27T04:51:39 * Naktibalda should tell something about people living at the bottom of Earth 2010-09-27T04:53:02 i guess nobody else thinks that's a desirable property then 2010-09-27T04:53:14 i would have thought maximising consistency would be desirable :\ 2010-09-27T04:53:27 even if it makes no difference for better bots 2010-09-27T04:54:12 i agree though that if on the same map switching positions causes a bot to lose 2010-09-27T04:54:14 something is wonky 2010-09-27T04:54:25 it's the bots that are wonky in a way 2010-09-27T04:54:32 if maps are symmetrical switching positions shouldn't matter 2010-09-27T04:54:39 but imo it's possible to eliminate that and it's really wouldn't be very computationally expensive 2010-09-27T04:54:52 bhasker: it's to do with the numbering of planets 2010-09-27T04:55:03 hmm but i don't use the planet numbering in any way 2010-09-27T04:55:05 at least that is my guess 2010-09-27T04:55:09 i am not sure why that should affect anything 2010-09-27T04:55:13 yes you do, you iterate through them 2010-09-27T04:55:15 my bot has the same problem on 44 2010-09-27T04:55:30 and dual bot picks differently on that 2010-09-27T04:55:37 it wouldn't if done my way 2010-09-27T04:55:43 hmm though i don't really use the ordering in any meaningful way 2010-09-27T04:55:47 i iterate but build my own queue 2010-09-27T04:55:56 so the actual physical order doesn't matter 2010-09-27T04:56:03 if you tie break, do you pick the first one you come across (or last, or any specific one for that matter)? 2010-09-27T04:56:05 i guess my bot just sucks on that map 2010-09-27T04:56:21 my queue is not based priorities 2010-09-27T04:56:39 it's possible for a bot to change it's moves currently because of that, changing the engine would stop that 2010-09-27T04:56:40 hmm though its possible that it may affect 2010-09-27T04:57:18 how often does the official server match up bots? 2010-09-27T04:57:21 my bot got one game 2010-09-27T04:57:26 and then nothing for the last few minutes 2010-09-27T04:57:40 Top 10 players: dmj111(3635), adumlah(3519), bix0r4ever(3505), Rexxar(3477), sequoh(3462), kcm1700(3454), davidjliu(3452), McLeopold(3419), sighbot(3375), vinx33(3369) 2010-09-27T04:57:46 few minutes or few hours? 2010-09-27T04:58:03 yeah, the server is struggling this competition, they're working on that 2010-09-27T04:58:23 1 game every 1 or 2 hours usually 2010-09-27T04:59:05 hmm seems to be faster than that 2010-09-27T04:59:13 i have 9 minutes between my games 2010-09-27T04:59:16 just got the second game 2010-09-27T04:59:34 oh,it's just a start 2010-09-27T04:59:41 my bot played 6 games in 12 hours 2010-09-27T04:59:56 what place? 2010-09-27T05:00:32 i just uploaded my first working bot to the official server had to make the code 2.5 python compatible 2010-09-27T05:01:08 http://ai-contest.com/profile.php?user_id=5822 2010-09-27T05:01:18 i wouldn't take rankings on the main site as very accurate at this point 2010-09-27T05:01:27 you just don't get enough games to generate anything reasonable 2010-09-27T05:01:52 the unofficial server anyway seems to have most of the good bots 2010-09-27T05:02:01 I think that bots after upload have a higher priority for first hour or something 2010-09-27T05:02:09 ooh k 2010-09-27T05:02:14 yea i just got my 3rd game 2010-09-27T05:03:26 *** naglijo has quit IRC (Ping timeout: 240 seconds) 2010-09-27T05:04:38 *** Palmik has quit IRC (Remote host closed the connection) 2010-09-27T05:04:54 i just wish people won't run 20 different variants of their bot on the unofficial server 2010-09-27T05:04:57 bhasker: your bot is pretty good 2010-09-27T05:05:01 *** shades has quit IRC (Ping timeout: 272 seconds) 2010-09-27T05:05:02 thanks 2010-09-27T05:05:10 been watching yours and adums 2010-09-27T05:05:17 mine? o_O 2010-09-27T05:05:32 wasn't yours on the top few on the unofficial 2010-09-27T05:05:35 sigh-timeout? 2010-09-27T05:05:38 isn't that yours? 2010-09-27T05:05:57 haha... yeah, although I haven't used sigh-timeout for a while 2010-09-27T05:06:06 that was when I was timing out all the time due to connection :P 2010-09-27T05:06:11 ah k 2010-09-27T05:06:18 i have improved my bot in the last two weeks 2010-09-27T05:06:26 it was quite dumb earlier, this is my best version yet 2010-09-27T05:06:30 seems to do decently 2010-09-27T05:06:45 one heuristic i need to build yet is when to acquire neutrals rather when not to 2010-09-27T05:06:49 my bot is too aggresive 2010-09-27T05:06:56 sometimes gets too behind on ships trying to acquire neutrals 2010-09-27T05:06:57 *** Palmik has joined #aichallenge 2010-09-27T05:07:04 that's always a hard choice 2010-09-27T05:07:17 but your bot uses quite sophisticated tactics 2010-09-27T05:07:31 not what I was expected when you said it was your first submission :P 2010-09-27T05:07:36 *** espes has quit IRC (Remote host closed the connection) 2010-09-27T05:07:41 haha no to the official server first one 2010-09-27T05:07:47 been trying it out on the unofficial 2010-09-27T05:07:58 different variants over last couple of weeks 2010-09-27T05:08:12 its ranked 70 now on the unofficial server 2010-09-27T05:08:12 yeah, fair enough 2010-09-27T05:08:28 http://72.44.46.68/getplayer?player=bhaskerv20.1 2010-09-27T05:08:41 *** aerique has quit IRC (Quit: ...) 2010-09-27T05:09:18 i lost quite a few games because i took neutrals too aggresively 2010-09-27T05:09:19 *** wvdschel has joined #aichallenge 2010-09-27T05:09:30 *** shades has joined #aichallenge 2010-09-27T05:10:03 *** sinan has joined #aichallenge 2010-09-27T05:10:13 adums bot is very interesting to watch, though right now it stomps over my bot in 50 turns:( 2010-09-27T05:11:04 it seems you've yet to win against my bot either :P 2010-09-27T05:11:11 which one is yours? 2010-09-27T05:11:25 not playing right now, my bots are all sigh-* 2010-09-27T05:11:34 but I've got logs from all the games my bot has played on tcp 2010-09-27T05:11:35 ah k 2010-09-27T05:11:48 well my earlier versions were rather dumb:-\ 2010-09-27T05:11:53 this seems to be the best one yet 2010-09-27T05:12:02 so it might do better against your bots now 2010-09-27T05:12:02 cool 2010-09-27T05:12:05 :) 2010-09-27T05:12:17 i had some basic bugs that i needed to work out 2010-09-27T05:12:18 I'll try to put one up in a couple of hours 2010-09-27T05:12:22 with how i handled fleet resolutions 2010-09-27T05:12:35 my bot would send out ships when it should not have 2010-09-27T05:12:39 and lose a planet 2010-09-27T05:12:41 etc 2010-09-27T05:12:46 yeah, I've had those problems too 2010-09-27T05:12:48 so took quite a few iterations to fix those bugs 2010-09-27T05:12:57 this one doesn't do any of that dumb crap 2010-09-27T05:13:13 :) 2010-09-27T05:13:27 though i still haven't gotten one bug worked out 2010-09-27T05:13:38 sending ships to capture neutral right after the enemy when my planet is closer 2010-09-27T05:13:46 for some reason my bot tends to launch the ships a bit too earl 2010-09-27T05:13:49 early* 2010-09-27T05:13:57 and ends up helping the enemy instead 2010-09-27T05:14:02 by wasting ships against the neutral:-\ 2010-09-27T05:15:45 I just explicitly forbid my bot from attacking neutrals closer to the enemy 2010-09-27T05:15:52 yea i need to do that 2010-09-27T05:15:57 though i have won matches 2010-09-27T05:15:59 when i do that too 2010-09-27T05:16:09 so i don't want it to be a strict heuristic 2010-09-27T05:16:14 rather strict command 2010-09-27T05:16:18 yeah 2010-09-27T05:16:20 want a more flexibile heuristic 2010-09-27T05:16:27 I wish that restriction was emergent from my heuristics too 2010-09-27T05:16:40 it shows that my current heuristics are not that good 2010-09-27T05:16:47 *** espes has joined #aichallenge 2010-09-27T05:16:55 but every time I try to mess with the my bot does worse 2010-09-27T05:17:06 heh i added a similar heuristic at start of game 2010-09-27T05:17:12 i restrict the visible space to my bot 2010-09-27T05:17:14 pingpong 2010-09-27T05:17:15 to 50% of the map 2010-09-27T05:17:18 hey bart 2010-09-27T05:17:27 and then expand it as the game progresses 2010-09-27T05:17:42 that way my bot doesn't do dumb stuff like sending ships across the map to some random planet 2010-09-27T05:17:53 bhasker: haven't tried anything like that myself 2010-09-27T05:18:03 try it out 2010-09-27T05:18:08 bhasker: i don't see why that wouldn't be emergent from your heuristic 2010-09-27T05:18:14 i tried that 2010-09-27T05:18:16 doesn't work:( 2010-09-27T05:18:16 *couldn't 2010-09-27T05:18:23 my heuristic always gets wonky 2010-09-27T05:18:32 because some stupid planet is sitting with 3-4 production 2010-09-27T05:18:39 with a small number of ships on it 2010-09-27T05:18:43 and my bot would just go after it 2010-09-27T05:18:49 even if it was right next to the enemy 2010-09-27T05:19:17 i tried various heuristics but my bot would always do something like that 2010-09-27T05:19:20 yeah, that's why my bot fails too... it under-estimates the strength of planets that can get easy reinforcements 2010-09-27T05:19:28 then you need to adjust the expected valueof the capture with parameters related to worstcase/expected case income unde retaliation and such 2010-09-27T05:19:46 yea i need to but restricting the search space works quite well too 2010-09-27T05:19:54 since most maps are symmetrical 2010-09-27T05:19:58 infact i lose due to being too cautious due to this and not capturing enough against more aggresive opponents 2010-09-27T05:20:12 happens to me at times 2010-09-27T05:20:25 http://72.44.46.68/getplayer?player=bhaskerv20.1 2010-09-27T05:20:28 yeah, atm my bot is quire defensive 2010-09-27T05:20:45 bart also i think your supply chain model needs a few changes 2010-09-27T05:20:53 adum's has perfected that model 2010-09-27T05:21:16 relly ? 2010-09-27T05:21:24 yes 2010-09-27T05:21:27 check his games out 2010-09-27T05:21:33 its quite interesting 2010-09-27T05:21:37 what his bot does 2010-09-27T05:21:42 i stopped using the hardcoded supply chain, and tried to make it emergent 2010-09-27T05:21:49 i implemented some of that stuff in my bot 2010-09-27T05:21:53 and its doing much better 2010-09-27T05:22:10 i think i understand what adum's bot does, its pretty similar to mine except his took it one step further 2010-09-27T05:22:12 bartwe_: how did that work out? 2010-09-27T05:22:24 adum's bot uses a concept that i call staging planet 2010-09-27T05:22:30 it funnels everything to a staging planet 2010-09-27T05:22:35 and always attacks from there 2010-09-27T05:22:37 except 2010-09-27T05:22:41 bhasker: mine does that too, but it is purely emergent 2010-09-27T05:22:46 the staging planet is not always its own planet 2010-09-27T05:22:55 he predicts which planet will be won 2010-09-27T05:22:56 sigh: i later found a bug in my statging code 2010-09-27T05:23:00 in the next few turns 2010-09-27T05:23:00 i can now make it stronger 2010-09-27T05:23:03 and uses that to stage 2010-09-27T05:23:07 and my bot does that too now 2010-09-27T05:23:15 it fails to take into account where the staging planet is going to be at the time of arrival 2010-09-27T05:23:18 so instead of reacting after capturing the planet 2010-09-27T05:23:27 its proactive 2010-09-27T05:23:46 adum's bot does it much better than mine 2010-09-27T05:23:53 and i don't funnel ships from one to the other 2010-09-27T05:23:59 i tend to send them straight 2010-09-27T05:24:11 been debating if a spanning tree is more efficient or not 2010-09-27T05:24:21 bartwe_ tried MST 2010-09-27T05:24:26 I don't think it is good 2010-09-27T05:24:29 yea i tried it 2010-09-27T05:24:31 didnt work out 2010-09-27T05:24:35 when you lose planets at back 2010-09-27T05:24:39 reversing the flows 2010-09-27T05:24:40 is a pain 2010-09-27T05:24:45 :P 2010-09-27T05:24:47 and you lose time 2010-09-27T05:24:55 with all your ships going in one direction or the other 2010-09-27T05:25:36 this game is a good example of my "staging": http://www.ai-contest.com/visualizer.php?game_id=4817924 2010-09-27T05:25:44 i still think some level of intermediate stops are superior over sending straight 2010-09-27T05:26:01 so this is where sigh i was saying your staging needs improvement 2010-09-27T05:26:06 you are staging to aplanet you control already 2010-09-27T05:26:10 its better to send it further ahead 2010-09-27T05:26:14 to a planet you will take control off 2010-09-27T05:26:14 yup, I know 2010-09-27T05:26:16 of* 2010-09-27T05:26:22 thats what adum's bot does 2010-09-27T05:26:25 I tried, it is hard... project for later 2010-09-27T05:26:29 the other thing i realized watching his bot 2010-09-27T05:26:30 bhasker: yeah i was planning to do that 2010-09-27T05:26:47 is he never waits to accumulate enough ships on one planet to capture a neutral 2010-09-27T05:26:55 he launches ships from closer planets as long as he has enough 2010-09-27T05:26:57 infact in a timestep different planets could be sending to different staging tactics due to the relativity of time/distance 2010-09-27T05:27:00 and i did that 2010-09-27T05:27:06 and it actually works quite well 2010-09-27T05:27:18 so i do something similar 2010-09-27T05:27:35 i sort my planet list by distance from destination and launch ships going by proximity 2010-09-27T05:27:53 bhasker: i would lauch sorted by expected value 2010-09-27T05:28:02 which may be suboptimal 2010-09-27T05:28:13 *** cmatei has quit IRC (Read error: No route to host) 2010-09-27T05:28:25 i use two queues 2010-09-27T05:28:33 one is list of potential targets 2010-09-27T05:28:38 and the other is the list of available ships 2010-09-27T05:28:39 bartwe_: expected value of what? 2010-09-27T05:28:51 i sort my available list everytime by distance from each target 2010-09-27T05:28:53 as i pick up targets 2010-09-27T05:29:05 expected value of a travel 2010-09-27T05:29:13 or a batch of travels actually 2010-09-27T05:29:41 http://72.44.46.68/canvas?game_id=184687 2010-09-27T05:29:50 bartwe_: batch of travels targeting the same planet? 2010-09-27T05:29:59 no, any batch of travels 2010-09-27T05:30:08 including future trvavels 2010-09-27T05:30:13 got it 2010-09-27T05:30:25 which is a bitch to calculate the budget requiriement for 2010-09-27T05:30:33 i never was very good at probability and expected values ! 2010-09-27T05:30:35 :-\ 2010-09-27T05:30:43 my heuristics are very simple :-\ 2010-09-27T05:30:51 bhasker: just do somethign so the bad < good < better 2010-09-27T05:31:01 bartwe_: but you still evaluation is a sum of per planet scores, right? 2010-09-27T05:31:14 yup i do something like that 2010-09-27T05:31:14 s/you still/your/ 2010-09-27T05:31:15 i now check dozends if attributes and future model 2010-09-27T05:31:20 and i'm down at lvl 50 atm 2010-09-27T05:31:43 mega1: no 2010-09-27T05:31:49 another game http://72.44.46.68/canvas?game_id=185057 for staging ahead of time 2010-09-27T05:32:23 i only model relative fitness with and without a batch of trvales 2010-09-27T05:33:26 I can't figure out how you decide to start attacking :D 2010-09-27T05:33:32 my bot? 2010-09-27T05:33:36 yeah 2010-09-27T05:33:53 as i said i increase the space ovetime 2010-09-27T05:34:00 so initially on some maps it won't even see the enemy 2010-09-27T05:34:10 as for attacking 2010-09-27T05:34:16 i have a concept that I call flow rate 2010-09-27T05:34:22 i use that to identify target planets 2010-09-27T05:34:36 and try to capture planets with highest flow rates 2010-09-27T05:34:39 *** sinan has quit IRC (Ping timeout: 276 seconds) 2010-09-27T05:35:03 flow rate is a property of the planet? 2010-09-27T05:35:10 and the distance from my planets 2010-09-27T05:35:15 basically something liek 2010-09-27T05:35:18 or depends on the state, is what I mean 2010-09-27T05:35:29 ok, I see 2010-09-27T05:35:32 (production/distance to planet)*(200-TurnNumber)-dest.NumShips 2010-09-27T05:35:46 yeah, I've been mucking around with something like that myself 2010-09-27T05:35:48 so it gives me a number as to how many ships that planet can provide 2010-09-27T05:35:51 over time 2010-09-27T05:35:51 looking at centrality measures 2010-09-27T05:35:52 to my planets 2010-09-27T05:36:00 and i sum that number to each of my planets 2010-09-27T05:36:05 and use that order my list of targets 2010-09-27T05:36:15 then i factor in ships around the planet/ships incoming etc 2010-09-27T05:36:21 and decide if i have enough ships to capture it etc 2010-09-27T05:36:31 I thinking looking to move 200 is problamatic 2010-09-27T05:36:43 and artificial, unless in endgame 2010-09-27T05:37:23 the 200 move limit doesn't matter until you are at least past 100, but more like 150 2010-09-27T05:37:25 well overall it works out quite well,except when sometimes in maps small planets which are closer 2010-09-27T05:37:30 cool 2010-09-27T05:37:37 have an artificially inflated flow rate 2010-09-27T05:37:48 so my bot would end up picking the wrong planet to capture 2010-09-27T05:37:54 yup 2010-09-27T05:37:57 *** dhartmei- has joined #aichallenge 2010-09-27T05:38:10 so i am still playing with my heuristic but for now i am sticking with this , till i can come up with a better way to select 2010-09-27T05:38:36 http://ai-contest.com/profile.php?user_id=5822 2010-09-27T05:39:12 bart what are you using as a heuristic to decide best neutrals to capture? 2010-09-27T05:39:13 huzzah i came up with something new to try :) 2010-09-27T05:39:50 does anyone know whose bots are these 9-/8- bots 2010-09-27T05:40:06 FlagCapper i think 2010-09-27T05:40:18 he has so many versions running 2010-09-27T05:40:33 i have a few usernames up 2010-09-27T05:40:43 it's kind of nice when you've made some changes to get a new elo 2010-09-27T05:40:52 i try to do it somewhat reservedly 2010-09-27T05:40:52 yea i do that too 2010-09-27T05:41:00 but not as much as these 8-/9- 2010-09-27T05:41:12 and i usually only run 1/2 bots at a time 2010-09-27T05:41:24 should have accounts and just let people reset their elo if they want 2010-09-27T05:41:35 but that involves effort 2010-09-27T05:41:51 also wouldn't allow people to imitate bartwe so easily 2010-09-27T05:42:05 haha 2010-09-27T05:42:29 apparently someone actually bothered aha 2010-09-27T05:42:55 bart which versions of your bots are currently active? 2010-09-27T05:43:01 i see an 8.2 2010-09-27T05:43:06 12.2 2010-09-27T05:43:39 i'm pretty sure it was 10 or 12 that was the imposter 2010-09-27T05:43:43 not sure if he is 12.2 or not 2010-09-27T05:43:53 heh 2010-09-27T05:44:37 how is this possible 2010-09-27T05:44:38 3 flavor 19h40m14s 19h40m14s 617 378 600 1 0% 755 0% 2010-09-27T05:44:42 hmm 2010-09-27T05:44:46 it ate up all the spaces 2010-09-27T05:51:32 another game showing the staging concept http://72.44.46.68/canvas?game_id=185274 2010-09-27T05:57:41 Top 10 players: dmj111(3644), adumlah(3520), bix0r4ever(3493), sequoh(3468), davidjliu(3464), kcm1700(3454), McLeopold(3442), Rexxar(3427), sighbot(3385), Zilog8(3361) 2010-09-27T05:58:40 *** elimisteve has joined #aichallenge 2010-09-27T06:11:19 *** iFire has joined #aichallenge 2010-09-27T06:12:41 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-09-27T06:16:33 *** iFire has joined #aichallenge 2010-09-27T06:26:42 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-09-27T06:42:37 *** iFire has joined #aichallenge 2010-09-27T06:44:02 *** HakanD has joined #aichallenge 2010-09-27T06:44:06 h 2010-09-27T06:44:07 h' 2010-09-27T06:44:11 hi 2010-09-27T06:44:54 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-09-27T06:45:06 hi, HakanD 2010-09-27T06:45:51 bah, changing keyboard layout is hard 2010-09-27T06:45:54 tons of typos 2010-09-27T06:46:03 :) 2010-09-27T06:46:22 *** iFire has joined #aichallenge 2010-09-27T06:48:29 *** iFire has joined #aichallenge 2010-09-27T06:50:26 *** iFire has joined #aichallenge 2010-09-27T06:54:32 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-09-27T06:55:54 *** iFire has joined #aichallenge 2010-09-27T06:57:41 Top 10 players: dmj111(3714), adumlah(3533), bix0r4ever(3502), sequoh(3486), davidjliu(3474), McLeopold(3462), kcm1700(3458), Rexxar(3440), sighbot(3398), Zilog8(3390) 2010-09-27T06:58:48 *** iFire has joined #aichallenge 2010-09-27T07:00:33 *** iFire has joined #aichallenge 2010-09-27T07:02:02 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-09-27T07:03:36 *** iFire has joined #aichallenge 2010-09-27T07:04:33 In response to the people that said I have too many versions of my bot running: I only run one at a time. 2010-09-27T07:04:59 one game per username? :P 2010-09-27T07:05:07 ../jks 2010-09-27T07:05:14 *** espes has quit IRC (Read error: Connection timed out) 2010-09-27T07:05:21 ;P 2010-09-27T07:05:37 *** espes has joined #aichallenge 2010-09-27T07:08:23 *** Cyndre_ has joined #aichallenge 2010-09-27T07:09:13 *** Cyndre has quit IRC (Read error: Connection reset by peer) 2010-09-27T07:11:39 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-09-27T07:14:52 *** BtbN has joined #aichallenge 2010-09-27T07:16:07 *** braddunbar has quit IRC (Ping timeout: 272 seconds) 2010-09-27T07:20:06 *** amstan has joined #aichallenge 2010-09-27T07:20:06 *** ChanServ sets mode: +o amstan 2010-09-27T07:20:34 *** Mathiasdm has quit IRC (Read error: Operation timed out) 2010-09-27T07:25:11 *** Naith has quit IRC (Quit: Naith) 2010-09-27T07:26:44 *** iFire has joined #aichallenge 2010-09-27T07:29:31 *** aerique has joined #aichallenge 2010-09-27T07:34:17 *** Mathiasdm has joined #aichallenge 2010-09-27T07:38:47 *** wvdschel has quit IRC (Quit: Leaving) 2010-09-27T07:57:41 Top 10 players: dmj111(3714), adumlah(3544), bix0r4ever(3494), sequoh(3484), davidjliu(3463), McLeopold(3461), Rexxar(3445), kcm1700(3442), sighbot(3422), savchenko(3414) 2010-09-27T07:59:28 woah, beat kcm1700: http://www.ai-contest.com/visualizer.php?game_id=4829190 2010-09-27T07:59:34 didn't expect that 2010-09-27T08:00:51 *** delt0r_ has joined #aichallenge 2010-09-27T08:02:16 *** delt0r___ has quit IRC (Ping timeout: 245 seconds) 2010-09-27T08:06:23 sigh: nice 2010-09-27T08:06:38 *** Arthur has quit IRC (Disconnected by services) 2010-09-27T08:07:04 meanwhile dmj is nearly 200 points above 2nd o_O 2010-09-27T08:07:53 *** Arthur has joined #aichallenge 2010-09-27T08:08:49 nice 2010-09-27T08:09:05 *** sinan has joined #aichallenge 2010-09-27T08:09:08 i have started some work on cleaning up and putting more strategy in 2010-09-27T08:09:16 cool 2010-09-27T08:09:22 gonna be a little bit before anything worth testing 2010-09-27T08:09:28 i should really be doing other work aha 2010-09-27T08:09:35 I started putting all my options into config 2010-09-27T08:09:43 current ranking on main site is laughable aha 2010-09-27T08:09:51 but will probably stay like that for the remainder of the week at least 2010-09-27T08:10:02 where are you? 2010-09-27T08:10:07 400's :P 2010-09-27T08:10:19 what happened? 2010-09-27T08:10:21 i put up a bot with lots of bugs, and cbf finding a better working copy 2010-09-27T08:10:37 haha... get version control 2010-09-27T08:10:37 and working on making things better the long way now 2010-09-27T08:10:38 *** Arthur has quit IRC (Disconnected by services) 2010-09-27T08:10:51 *** Arthur_ has joined #aichallenge 2010-09-27T08:11:46 having said that, it's not "that" buggy 2010-09-27T08:11:54 the bots on the server are certainly getting better at that level 2010-09-27T08:12:13 yeah, level keeps getting higher 2010-09-27T08:13:09 i still think top 10 finish is doable if we get things right 2010-09-27T08:13:28 :) 2010-09-27T08:13:33 but it's certainly going to be more competitive than last time 2010-09-27T08:13:33 hm, is it possible to assign IDs to fleets? 2010-09-27T08:13:39 or is the order random? 2010-09-27T08:13:50 BtbN: for where? 2010-09-27T08:13:53 BtbN: order is whatever the server gives you 2010-09-27T08:14:01 you can't assume anything about it 2010-09-27T08:14:05 to store if i already reacted to that fleet 2010-09-27T08:14:23 i store my own fleets, only adding new ones as i pass through the game state string 2010-09-27T08:14:26 BtbN: just check for fleets which have length-turnsremaining = 1 2010-09-27T08:14:30 easy 2010-09-27T08:14:54 Nah, not in my case 2010-09-27T08:16:13 why can't you do that? 2010-09-27T08:16:38 cause the AI might wait for the planet to have enough resources 2010-09-27T08:16:50 and some other cases where is pauses 2010-09-27T08:17:02 *** sinan has quit IRC (Ping timeout: 240 seconds) 2010-09-27T08:17:08 assign ids internally, create new ids for new fleets (which you find with the equation I gave you) 2010-09-27T08:17:25 ah, yeah. Nice Idea 2010-09-27T08:17:48 you only care about where a fleet is going, when it gets there and who owns it 2010-09-27T08:17:55 in particular, not where it's from 2010-09-27T08:18:16 antimatroid: I thought so, but now I'm now starting to care where they are from :P 2010-09-27T08:18:37 to send some hate back? 2010-09-27T08:18:39 meh, i still think it's irellevant 2010-09-27T08:18:57 it's relevant for the visualizer 2010-09-27T08:18:59 except handling it when you se it 2010-09-27T08:20:06 BtbN: to predict opponent behaviour 2010-09-27T08:20:31 sigh: you can handle that you you see the fleet 2010-09-27T08:20:49 *** dmj111 is now known as dmj111_ 2010-09-27T08:22:06 antimatroid: yeah, I don't store the sender during my main processing 2010-09-27T08:23:07 *** zaphus has joined #aichallenge 2010-09-27T08:31:05 *** kjeldahl_ has joined #aichallenge 2010-09-27T08:31:32 *** sinan has joined #aichallenge 2010-09-27T08:34:02 *** kjeldahl has quit IRC (Ping timeout: 240 seconds) 2010-09-27T08:35:38 *** krokokrusa has joined #aichallenge 2010-09-27T08:39:27 *** HakanD has quit IRC (Read error: Connection reset by peer) 2010-09-27T08:39:32 *** HakanD_ has joined #aichallenge 2010-09-27T08:40:50 *** aerique has quit IRC (Quit: ...) 2010-09-27T08:47:19 *** ioasdjfo has joined #aichallenge 2010-09-27T08:55:54 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-09-27T08:57:42 Top 10 players: dmj111(3704), adumlah(3541), bix0r4ever(3480), sighbot(3469), davidjliu(3460), sequoh(3447), kcm1700(3439), McLeopold(3436), savchenko(3421), Rexxar(3403) 2010-09-27T08:58:12 oh wow, huge rating jump 2010-09-27T08:58:29 *** iFire has joined #aichallenge 2010-09-27T08:58:45 hehe 2010-09-27T08:59:10 turn out I beat McLeopold, Rexxar and kcm1700 in the last few games 2010-09-27T08:59:34 nice, well done on the ranking ;) 2010-09-27T08:59:45 :) 2010-09-27T09:01:39 http://72.44.46.68/canvas?game_id=187356 Strange game 2010-09-27T09:01:54 wtf doesnt my bot attack? 2010-09-27T09:02:26 rather strange concept, backing up max growth rather than closest enemy planets 2010-09-27T09:02:28 bug in bot 2010-09-27T09:02:41 *** Naith has joined #aichallenge 2010-09-27T09:02:46 *** otherAntimatroid has left #aichallenge 2010-09-27T09:02:52 *** zl has quit IRC (Quit: Page closed) 2010-09-27T09:03:27 BtbN: which bot is yours? 2010-09-27T09:03:31 Oromit 2010-09-27T09:04:09 is that redistribution strategy intentional? 2010-09-27T09:04:13 probably you compare a count of your ships against a enemy and it isn't enough to capture planet 2010-09-27T09:04:36 it should see that it has more than enough ships to take one of the two enemy planets 2010-09-27T09:04:46 are you defending too much? 2010-09-27T09:04:54 No Attack-code in there atm :D 2010-09-27T09:05:05 just this simple one 2010-09-27T09:05:14 if more ships than enemy + growth rate, attack 2010-09-27T09:05:43 oh well, should be easy to track down then 2010-09-27T09:05:50 just log what you bot is thinking :) 2010-09-27T09:06:13 i should compile the code-changes I do :/ 2010-09-27T09:06:17 *** troutwine has joined #aichallenge 2010-09-27T09:07:13 someone posted this in the forums: http://ai-contest.com/visualizer.php?game_id=4811861 2010-09-27T09:07:17 cool game! 2010-09-27T09:07:55 yep, here's for hoping games all end up being a lot like that near the top 2010-09-27T09:08:03 if not even better fought 2010-09-27T09:08:28 i hope the code for this will get published when the war is over 2010-09-27T09:08:31 :) 2010-09-27T09:08:35 it will 2010-09-27T09:08:49 and i hope it isn't Haskel 2010-09-27T09:08:55 haha 2010-09-27T09:09:01 someone should compile a list of cool games 2010-09-27T09:09:02 meh, i need to learn it sometime 2010-09-27T09:09:11 for ai? 2010-09-27T09:09:15 jeff cameron has a list somewhere 2010-09-27T09:09:27 tron was awesome for a simpler game 2010-09-27T09:09:31 no as in cool matches in the contest 2010-09-27T09:09:41 oh, yep 2010-09-27T09:09:43 like the one posted :P 2010-09-27T09:10:05 I would do it, but I don't watch that many games 2010-09-27T09:10:17 yeah neither, except my own 2010-09-27T09:10:29 for debugging purposes mostly too 2010-09-27T09:10:35 or to see how well i ruined someone 2010-09-27T09:10:37 yeah, me too 2010-09-27T09:10:38 haha 2010-09-27T09:10:53 it's satisfying to beat someone in under 70 moves 2010-09-27T09:11:22 just keep resubmitting your bot, and playing people ~1000 :P 2010-09-27T09:11:29 atm my bot wins most of it games in 200 2010-09-27T09:11:35 as it does a "growth-race" :/ 2010-09-27T09:11:40 in 200? 2010-09-27T09:11:45 ah, 200 moves 2010-09-27T09:12:00 btw, there some good different modes for galcon on iphone app 2010-09-27T09:12:44 BtbN: well, the player losing the growth race should do something :) 2010-09-27T09:12:53 changing fleet destinations on the move is funny 2010-09-27T09:13:09 sigh: yeah, obviously :D 2010-09-27T09:15:01 *** iouri_ has quit IRC (Quit: Page closed) 2010-09-27T09:16:06 just put in my first 4 dimensional vector :| 2010-09-27T09:16:11 i should really just use my tensor class 2010-09-27T09:16:55 don't forget about quaternions 2010-09-27T09:17:22 you are using tensors in your AI? o_O 2010-09-27T09:18:02 tensor in that it's just like an nD matrix 2010-09-27T09:18:20 so say i take my 4d tensor, it just goes into that like there are 4 dimensions 2010-09-27T09:18:35 then has a vector specifying the "length" of each dimension 2010-09-27T09:18:41 and all values are stored on a 1d vector 2010-09-27T09:18:51 use it for a game theory library i'm working on 2010-09-27T09:19:18 antimatroid: I would to speak theory with you 2010-09-27T09:19:29 ...? 2010-09-27T09:19:29 So far I saw different approaches. 2010-09-27T09:20:04 i'm not very knowledgable, i just think of stuff then start searching for anything relevant most of the time 2010-09-27T09:20:05 1 - is to have an heuristic to attack and defend against fleets. This heuristic is widely adaptable, but I didn't try to put it far. 2010-09-27T09:20:29 2 - is to use a graph 2010-09-27T09:20:45 i haven't not thought about it as a graph in a long time 2010-09-27T09:20:51 ai algorithms are often dealing with graphs 2010-09-27T09:21:25 are you developping a matricial approach ? 2010-09-27T09:21:26 although i don't store it as such, as you already know every node is connected to every other node, just store distances 2010-09-27T09:21:41 not really, just useful way to think about it 2010-09-27T09:21:42 antimatroid: true 2010-09-27T09:22:27 the number of path is about (n-1)! if you consider a path can be used in the two directions 2010-09-27T09:22:54 actually exactly (n-1)! 2010-09-27T09:23:07 i haven't really made much use of graph theory algorithms, but i still find it the best way to think about the problem 2010-09-27T09:23:48 Do you actually think that the problem is "path finding" problem ? 2010-09-27T09:24:20 The matter is that this kind of graph doesn't fit with collaborative work between planets. 2010-09-27T09:24:20 http://ai-contest.com/profile.php?user_id=9494 he realy shoul fix his bot 2010-09-27T09:25:05 it's not really a path finding problem 2010-09-27T09:25:23 What's the problem for you ? 2010-09-27T09:25:27 but the problem is isomorphic to a graph theory one, so it's nice to just think of it in terms that I am used to 2010-09-27T09:25:33 to win? :P 2010-09-27T09:26:14 antimatroid: yes, but you win against someone else. The problem is to have an ai which can adapt to a broad quantity of situations. 2010-09-27T09:26:38 yes, and that can all be encapsulated into the objective of winning :P 2010-09-27T09:27:10 antimatroid: That's what I thought. 2010-09-27T09:27:38 But actually, if you see it as a graph problem, it seems rather difficult because the complexity is far too big. 2010-09-27T09:27:43 you might want to predict future states of planets based on known fleets if you aren't already 2010-09-27T09:27:58 antimatroid: I already tried this approach. 2010-09-27T09:28:08 It doesn't work perfectly. 2010-09-27T09:28:12 people have found it quite successful 2010-09-27T09:28:17 myself included 2010-09-27T09:28:28 Yes, but you cannot use intermediate states. 2010-09-27T09:28:38 intermediate? 2010-09-27T09:28:43 eg, a planet is neutral with one ship. 2010-09-27T09:28:45 it's a discrete game 2010-09-27T09:29:02 what about a neutral planet with one ship? 2010-09-27T09:29:06 http://72.44.46.68/canvas?game_id=187646 seems like both bots decided that map sucks :D 2010-09-27T09:29:31 Would you like to wait for your opponent to finish the work or get directly to the last neutral ship and keep the planet yours when the opponent already destroyed so many of his bots ? 2010-09-27T09:29:33 i think they could have taken that 13 planet couldn't they? 2010-09-27T09:29:56 depends on which generates you the most growth 2010-09-27T09:29:56 then they'd loose their home-planet, of the other bot is good 2010-09-27T09:30:20 no, i mean i think they could do that without sending so many that the other person could take them 2010-09-27T09:30:32 yeah, the better bots seem to determine how many ships they can use, and take over neutrals 2010-09-27T09:30:42 they are distanced like 4 apart yeah? with 5 growth, that's 20 growth generated in the time to send a fleet there 2010-09-27T09:31:03 Even they send just one more ship than needed to capture a planet. 2010-09-27T09:31:19 You just have to wait one elapsed round and take the planet just after. 2010-09-27T09:31:22 *** troutwine has quit IRC (Ping timeout: 264 seconds) 2010-09-27T09:31:35 yes, mine does that when it decides it's optimal to do so 2010-09-27T09:31:45 you send the the planets closer, you can easily defend if your opponent tried to take it 2010-09-27T09:32:00 *closer to you 2010-09-27T09:32:06 i don't even consider neutrals closer to them 2010-09-27T09:32:15 antimatroid: me neither 2010-09-27T09:32:16 although people have had a lot of success in predicting where you will move 2010-09-27T09:32:18 near the centre 2010-09-27T09:32:37 have you seen some of the dmj111 games ? 2010-09-27T09:32:52 yeah, haven't paid too much attention to them yet though 2010-09-27T09:33:05 i'm undecided on how much to redistribute and when 2010-09-27T09:33:07 *** mightybyte has joined #aichallenge 2010-09-27T09:33:07 he surround the enemy planet. Just take a planet far from the enemy but near of his planets. 2010-09-27T09:33:32 yeah, i wont touch an enemy planet from one of mine unless it's that planets closest enemy 2010-09-27T09:33:44 It's a perfect move cause it draws a circle around the target, then he can use many ships to attack the target and send just the appropriate quantity. 2010-09-27T09:33:52 although, that isn't optimal 2010-09-27T09:34:05 consider when the enemy planet is on either side of you, you may want to take the one further away 2010-09-27T09:34:14 enemy planets* 2010-09-27T09:35:12 the dmj111 way seems to be an argumentation in favor of quick victories. 2010-09-27T09:35:52 that could just be the result of his bot being much better than other current bots :p 2010-09-27T09:36:23 *** zl has joined #aichallenge 2010-09-27T09:37:03 http://ai-contest.com/visualizer.php?game_id=4810422 2010-09-27T09:37:18 IMO this is definitely a graph oriented approach. 2010-09-27T09:37:37 You can see that he's looking to the better path in the graph to kill the enemy 2010-09-27T09:37:57 *** ioasdjfo has quit IRC (Quit: Page closed) 2010-09-27T09:38:55 i dunno about that 2010-09-27T09:39:48 no i was mistaken 2010-09-27T09:40:05 the strategy is to send the more sheep possible to the center planets and keep them. 2010-09-27T09:40:15 After that he send many ships on the biggest planets. 2010-09-27T09:40:29 Path finding algorithm ? 2010-09-27T09:40:46 It seems IA doesn't do well the job, you have to program the behaviour your self. 2010-09-27T09:42:19 *** Nuriaion has joined #aichallenge 2010-09-27T09:43:16 http://ai-contest.com/visualizer.php?game_id=4824268 I dont understand quite well this kind of games. 2010-09-27T09:45:28 that game is easy to understand, my bot is an idiot and doesn't have any clue that it is behind (seriously) 2010-09-27T09:45:51 so it just tries to protect its planets 2010-09-27T09:45:57 ;) 2010-09-27T09:45:59 I think that it was easier to start from beginning. top bots weren't so good and I didn't feel so far behind 2010-09-27T09:46:20 agreed 2010-09-27T09:46:36 i saw the orginal premature post on reddit and was like zomg there goes my uni marks 2010-09-27T09:46:47 :) 2010-09-27T09:46:56 i also made a bot for tron, which was easier 2010-09-27T09:47:00 although awesome 2010-09-27T09:47:43 *** aerique has joined #aichallenge 2010-09-27T09:49:32 *** Appleman1234 has quit IRC (Ping timeout: 240 seconds) 2010-09-27T09:50:31 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-09-27T09:53:40 *** amstan has quit IRC (Remote host closed the connection) 2010-09-27T09:57:42 Top 10 players: dmj111(3699), adumlah(3535), bix0r4ever(3474), sighbot(3460), davidjliu(3460), sequoh(3439), kcm1700(3433), McLeopold(3427), user42(3364), Rexxar(3359) 2010-09-27T09:59:59 *** Nuriaion has quit IRC (Quit: Nuriaion) 2010-09-27T10:01:25 *** keyne has joined #aichallenge 2010-09-27T10:04:09 *** Olathe has joined #aichallenge 2010-09-27T10:08:02 *** iFire has joined #aichallenge 2010-09-27T10:15:29 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-09-27T10:18:26 dmj111 is first since a long time. 2010-09-27T10:18:39 so what's the quickest path to victory ? 2010-09-27T10:18:54 Define a strategy : less bad choice ? 2010-09-27T10:18:58 better choice ? 2010-09-27T10:19:20 no choice ? 2010-09-27T10:20:23 quickest path to create the best bot? or to win a single game? 2010-09-27T10:22:26 no to win a single game I mean 2010-09-27T10:23:01 is it a good idea for you to focus on the growth rate ? 2010-09-27T10:23:22 try to make your opponent growth rate smaller while you are trying to make yours higher ? 2010-09-27T10:23:51 I'm talking about choice you know, it 's rather difficult to have an idea of the criteria used to be good in the game. 2010-09-27T10:24:06 just don't forget that if you have +10 grwoth_rate but opponent has +200 ships, he can take over your planets fast 2010-09-27T10:24:41 Cause even if you define such a simple criteria, the choices are enormous. 2010-09-27T10:24:56 so which planet could you attack : neutral, enemy ? 2010-09-27T10:25:06 enemy :) 2010-09-27T10:25:13 fast growth rate, low num ship ? 2010-09-27T10:25:37 *** sinan has quit IRC (Ping timeout: 245 seconds) 2010-09-27T10:25:49 so let's take the enemy : quick victory, slow victory ? intermediate planets to conquer ? 2010-09-27T10:25:56 who knows ? 2010-09-27T10:25:59 :) 2010-09-27T10:26:43 thats the thing that makes this challenge interesting (: 2010-09-27T10:27:09 which theoretical arsenal would you use ? graph theory ? 2010-09-27T10:28:14 other ? 2010-09-27T10:28:35 Neural Network are said not to be good. I would be doubtfull since it requires good variables to use. 2010-09-27T10:28:40 *** Arthur_ has quit IRC (Ping timeout: 272 seconds) 2010-09-27T10:29:33 another question : 2010-09-27T10:29:41 are the default bots in the challenge ? 2010-09-27T10:30:04 maybe 2010-09-27T10:30:13 I would like to see a RageBot in the ranking 2010-09-27T10:30:14 *** Arthur_ has joined #aichallenge 2010-09-27T10:33:49 Someone does know its ranking ? 2010-09-27T10:34:32 *** troutwine has joined #aichallenge 2010-09-27T10:39:58 http://72.44.46.68/getratings 2010-09-27T10:40:04 RageBot has 28% win rate 2010-09-27T10:41:38 *** sigh has quit IRC (Remote host closed the connection) 2010-09-27T10:43:25 *** lotek93 has quit IRC (Ping timeout: 252 seconds) 2010-09-27T10:44:03 *** sinan has joined #aichallenge 2010-09-27T10:46:44 *** Cyndre_ has quit IRC (Quit: Leaving) 2010-09-27T10:49:40 *** Cyndre has joined #aichallenge 2010-09-27T10:52:50 *** di has quit IRC (Quit: Page closed) 2010-09-27T10:57:43 Top 10 players: dmj111(3697), adumlah(3538), bix0r4ever(3472), sighbot(3458), davidjliu(3453), kcm1700(3448), sequoh(3437), McLeopold(3421), delineate(3404), Rexxar(3364) 2010-09-27T11:04:54 *** u1tr0n has joined #aichallenge 2010-09-27T11:07:29 *** adesst has joined #aichallenge 2010-09-27T11:07:53 *** zaphus has quit IRC (Quit: Page closed) 2010-09-27T11:09:00 hello, would somebody kindly help me pls, my id is http://ai-contest.com/profile.php?user_id=8108, and am having problem with latest upload bot. MyBot runs well in my local computer, but now is losing at the 1st turn. 2010-09-27T11:09:34 is it compatible with python 2.5 2010-09-27T11:10:13 the latest submission started at Sep 27th 08:15:13, Sep 27th 09:52:55,Sep 27th 09:59:30 and yet i have not received any email that says "Compile Success" 2010-09-27T11:10:30 uhmm i am currently using 2.6 2010-09-27T11:12:39 I think that it should be possible to add syntax checking for interpreted languages to compile_anything.py 2010-09-27T11:12:40 @Naktibalda, thx for hint me, let me do a try, to run it using 2.5 2010-09-27T11:12:42 adesst: I'm sorry Dave, I cannot Naktibalda,. 2010-09-27T11:13:05 Naktibalda, thx for hint me, let me do a try, to run it using 2.5 2010-09-27T11:13:40 will back in 5 2010-09-27T11:13:51 *** HakanD_ has quit IRC (Ping timeout: 240 seconds) 2010-09-27T11:14:29 *** Itkovian has quit IRC (Read error: Operation timed out) 2010-09-27T11:14:46 *** iFire has joined #aichallenge 2010-09-27T11:18:54 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-09-27T11:24:15 *** Gudradain has joined #aichallenge 2010-09-27T11:24:31 hi 2010-09-27T11:25:34 *** iFire has joined #aichallenge 2010-09-27T11:27:02 Many thx Naktibalda, i found that "except Exception as e" syntax, is not compatible with 2.5 2010-09-27T11:27:02 will do another upload 2010-09-27T11:27:50 *** HakanD_ has joined #aichallenge 2010-09-27T11:32:09 hurray! my bot is up and running, and again thx Naktibalda for being so helpful 2010-09-27T11:32:59 especially considering the fact that Naktibalda himself codes in PHP :) 2010-09-27T11:33:47 * Naktibalda is trying to remember versions of all programming languages in contest server 2010-09-27T11:34:33 i too code php for everyday uses, and am learning python 2010-09-27T11:35:03 *** HakanD_ has quit IRC (Ping timeout: 240 seconds) 2010-09-27T11:35:40 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-09-27T11:35:46 ahhh just checked and found out that php is now supported, coolness 2010-09-27T11:36:13 but i'll stick with python :P 2010-09-27T11:39:32 *** amstan has joined #aichallenge 2010-09-27T11:39:33 *** ChanServ sets mode: +o amstan 2010-09-27T11:39:57 contestbot: topic remove 5 2010-09-27T11:39:57 *** contestbot changes topic to "Official Google AI Challenge: http://ai-contest.com/ || Channel Logs: http://contestbot.hypertriangle.com/ || No stupid questions. Don't ask to ask. || Developers wanted, contact j3camero. || Latest Idea, decentralization: http://tinyurl.com/decentralizedai" 2010-09-27T11:40:03 contestbot: topic remove 5 2010-09-27T11:40:04 *** contestbot changes topic to "Official Google AI Challenge: http://ai-contest.com/ || Channel Logs: http://contestbot.hypertriangle.com/ || No stupid questions. Don't ask to ask. || Developers wanted, contact j3camero." 2010-09-27T11:40:52 * rogue780 is pondering abandoning java and writing a php or python bot 2010-09-27T11:41:08 rogue780: python ftw 2010-09-27T11:41:46 python is pretty amazing for rapid development 2010-09-27T11:44:08 I prefer C# for rapid stuff 2010-09-27T11:44:16 Mostly because of the nice IDE 2010-09-27T11:44:40 BtbN: ide which only works on windows, meh 2010-09-27T11:45:11 MonoDevelop and SharpDevelop run on many OSes 2010-09-27T11:45:21 And they get quite close to VisualStudio 2010-09-27T11:48:15 *** HakanD_ has joined #aichallenge 2010-09-27T11:49:09 which user & pass i need to send to TCP server ? 2010-09-27T11:49:24 u1tr0n: make up a new one 2010-09-27T11:49:34 where ? 2010-09-27T11:49:42 u1tr0n: probably the first time 2010-09-27T11:49:44 then it remembers it 2010-09-27T11:49:45 u1tr0n: as long as you use the same username/password consistently it will just treat it as a single bot 2010-09-27T11:49:58 u1tr0n: you don't have to register, that is 2010-09-27T11:50:01 ok, tnx 2010-09-27T11:51:18 *** adesst has quit IRC (Ping timeout: 265 seconds) 2010-09-27T11:54:21 *** dassouki has joined #aichallenge 2010-09-27T11:54:40 *** HakanD_ has quit IRC (Ping timeout: 240 seconds) 2010-09-27T11:57:17 *** sinan has quit IRC (Ping timeout: 245 seconds) 2010-09-27T11:57:43 Top 10 players: dmj111(3695), adumlah(3532), bix0r4ever(3476), sequoh(3455), kcm1700(3443), davidjliu(3440), sighbot(3429), delineate(3411), McLeopold(3408), Rexxar(3359) 2010-09-27T11:59:10 *** adesst has joined #aichallenge 2010-09-27T12:01:54 *** dassouki has quit IRC (Quit: leaving) 2010-09-27T12:05:07 *** iFire has joined #aichallenge 2010-09-27T12:07:52 *** bhasker has quit IRC (Quit: bhasker) 2010-09-27T12:15:06 *** zl has quit IRC (Ping timeout: 252 seconds) 2010-09-27T12:16:10 use netbeans as your python IDE 2010-09-27T12:16:14 oy maybe eclipse 2010-09-27T12:20:11 nite all 2010-09-27T12:20:14 *** adesst has quit IRC (Quit: HydraIRC -> http://www.hydrairc.com <- Now with extra fish!) 2010-09-27T12:23:01 *** BtbN has quit IRC (Quit: Verlassend) 2010-09-27T12:25:26 *** bhasker has joined #aichallenge 2010-09-27T12:26:25 *** Bobng has joined #aichallenge 2010-09-27T12:28:50 *** aerique has quit IRC (Quit: ...) 2010-09-27T12:35:23 *** HakanD has joined #aichallenge 2010-09-27T12:41:26 *** dhartmei- has quit IRC (Quit: .) 2010-09-27T12:43:34 what is "=invalid input" ? 2010-09-27T12:45:52 negative number ships, more ships than are in planet, you don't own source planet, etc 2010-09-27T12:54:35 hello, can we join the challenge with a team from the same college ? 2010-09-27T12:57:36 *** coventry has joined #aichallenge 2010-09-27T12:57:44 Top 10 players: dmj111(3703), adumlah(3538), bix0r4ever(3480), sequoh(3460), kcm1700(3448), davidjliu(3447), delineate(3415), sighbot(3415), McLeopold(3402), Rexxar(3402) 2010-09-27T12:58:09 *** wh1teside has quit IRC (Ping timeout: 272 seconds) 2010-09-27T12:59:50 I'm really struggling to understand the principles behind the dramatic reversal in this game between McLeopold and dmj111. At move 56, McLeopold seems to be ahead in every way, then dmj111 stages this crazy invasion and wipes the board with him. http://ai-contest.com/visualizer.php?game_id=4811861 2010-09-27T13:01:11 McLeopold wasted too many ships capturing neutrals around turn 56 2010-09-27T13:01:53 notice turn 57; as soon as the 72-fleet to the 71-size neutral shows up, dmj counterattacks 2010-09-27T13:06:07 At turn 58, McLeopold had 922 ships (ignoring the fleet of 69 committed to the remote neutral.) dmj had 960. That's a 4-turn lead, given McLeopold's excess productive capacity of 10 ships per turn. He's already made the deficit up three turns before dmj's first attack fleet lands. 2010-09-27T13:06:57 I agree that McLeopold's committment to taking the neutrals is what triggered the attack from dmj, though. 2010-09-27T13:10:49 *** amstan has quit IRC (Remote host closed the connection) 2010-09-27T13:11:14 But at turn 60, McLeopold's subsequent commitment of 83- and 91-ship fleets to other neutrals may have been the crucial blunder. Thanks for pointing that out. 2010-09-27T13:12:57 *** coventry has left #aichallenge ("ERC Version 5.3 (IRC client for Emacs)") 2010-09-27T13:14:02 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716]) 2010-09-27T13:15:16 *** sinan has joined #aichallenge 2010-09-27T13:15:44 *** Gudradain has quit IRC (Quit: Page closed) 2010-09-27T13:16:31 *** Olathe has quit IRC (Ping timeout: 272 seconds) 2010-09-27T13:21:20 *** amstan has joined #aichallenge 2010-09-27T13:21:20 *** ChanServ sets mode: +o amstan 2010-09-27T13:28:04 *** justin_pdx has joined #aichallenge 2010-09-27T13:28:33 *** Migi32 has joined #aichallenge 2010-09-27T13:28:40 *** keyne has quit IRC (Quit: Inkl Web IRC :: http://barosl.com/webirc/) 2010-09-27T13:29:10 can I rename MyBot.cc and PlanetWars.cc? (or just change their extension to .cpp?) 2010-09-27T13:29:21 NetBeans is not working well with .cc file extensions 2010-09-27T13:31:15 .cpp is okay 2010-09-27T13:32:00 you can rename/delete/add any source files you like, but there has to be a MyBot.[extension] 2010-09-27T13:32:29 idk 2010-09-27T13:32:31 contestbot: repo 2010-09-27T13:32:32 amstan: repo = http://code.google.com/p/ai-contest/ 2010-09-27T13:33:48 yep, you can use cpp 2010-09-27T13:33:49 http://code.google.com/p/ai-contest/source/browse/trunk/planet_wars/backend/compile_anything.py 2010-09-27T13:36:28 *** nullkuhl has quit IRC (Ping timeout: 245 seconds) 2010-09-27T13:44:23 *** krokokrusa has quit IRC (Ping timeout: 272 seconds) 2010-09-27T13:46:50 *** amstan has quit IRC (Ping timeout: 265 seconds) 2010-09-27T13:51:17 *** troutwine has quit IRC (Ping timeout: 276 seconds) 2010-09-27T13:51:18 *** kjeldahl_ has quit IRC (Remote host closed the connection) 2010-09-27T13:57:45 Top 10 players: dmj111(3700), adumlah(3525), bix0r4ever(3466), kcm1700(3465), sequoh(3462), Titannet(3459), davidjliu(3432), delineate(3426), sighbot(3413), Rexxar(3405) 2010-09-27T13:59:01 dmj is on fire 2010-09-27T13:59:20 or is there ELO inflation? 2010-09-27T13:59:28 *** keynes has joined #aichallenge 2010-09-27T14:00:20 stupid question, that's not possible if nobody leaves the ELO listing 2010-09-27T14:00:56 anyone using dvorak keyboard layout here? 2010-09-27T14:01:22 I used Dvorak for ~1 month, why? 2010-09-27T14:01:47 thinking about switching to it (: 2010-09-27T14:02:07 I'd say do it, it's fun :D 2010-09-27T14:02:34 but be prepared to type retardedly slow for a while 2010-09-27T14:02:46 yep im aware of that (: 2010-09-27T14:03:20 especially while coding 2010-09-27T14:03:59 *** troutwine has joined #aichallenge 2010-09-27T14:04:05 but while learning it might surprise you how deep in our subconscious typing skills are baked 2010-09-27T14:04:18 best time to change is when doing extensive coding without a strict timeline i think 2010-09-27T14:04:21 which is now for me 2010-09-27T14:04:46 http://play.typeracer.com/ <- I used this 2010-09-27T14:05:27 also, if you'd use it to program a lot I'd suggest programmer dvorak 2010-09-27T14:05:46 basically the same except (){}[]:,; etc are all a lot more accessibleµ 2010-09-27T14:06:47 yep saw that on wiki 2010-09-27T14:06:59 Thanks for the info 2010-09-27T14:09:23 *** sinan has quit IRC (Ping timeout: 245 seconds) 2010-09-27T14:09:40 woah, it's been ages since I last typed on Dvorak, yet I can still do it on a somewhat decent speed 2010-09-27T14:09:51 *** hellman has quit IRC (Remote host closed the connection) 2010-09-27T14:11:37 *** keynes has quit IRC (Ping timeout: 272 seconds) 2010-09-27T14:12:59 im changing all the things about my computer life, so i ll probably try dvorak as well 2010-09-27T14:13:14 I just need some stickers for keyboard (: 2010-09-27T14:13:22 HakanD: I use dvorak 2010-09-27T14:14:31 mightybyte, wanna share your thoughts about changing from qwerty to dvorak- 2010-09-27T14:14:51 Getting to a 60wpm didn't take long. 2010-09-27T14:15:05 Getting up to 80-90 took much longer. 2010-09-27T14:15:20 And I still haven't reached the speed that I used to type with qwerty. 2010-09-27T14:15:42 But that speed took a LONG time to reach on qwerty. 2010-09-27T14:16:46 how long have you been using dvorak, mightybyte? 2010-09-27T14:16:52 So it worths investing that amount of time? 2010-09-27T14:16:54 i have heard that most people never reach their old qwerty speeds when they switch to dvorak, but that's not the point anyway 2010-09-27T14:16:58 Probably 4-5 years. 2010-09-27T14:17:14 jmcarthur: Well, that was one of the main reasons I tried it. 2010-09-27T14:17:30 I find that typing speed is just not something thats holding me back 2010-09-27T14:17:31 HakanD: That depends on your motivation. 2010-09-27T14:17:34 the only reason i've ever considered it is to prevent strain on my hands 2010-09-27T14:17:44 I've never had problems with strain. 2010-09-27T14:17:55 also HakanD, I'd say NO stickers. NO cheat sheets. I know it's annoying to memorize the keys but just do it. You'll learn much quicker afterwards 2010-09-27T14:18:01 nor i, but i know people who have, and i don't want that to happen to me 2010-09-27T14:18:12 actually, i have, but it's emacs' fault 2010-09-27T14:18:30 Migi32: I'll second that. If you're a hunt-and-peck qwerty guy, then switching to dvorak with no keyboard modifications would probably be a big benefit. 2010-09-27T14:18:31 Only reason im thinking about switching is the health reasons 2010-09-27T14:18:37 emacs has a short cut for that ;) 2010-09-27T14:18:45 a quick remapping of the capslock key to control was all i needed to fix my problem 2010-09-27T14:19:17 If I was going for speed, I don't think dvorak is the way to do it. I'd probably come up with a custom keyboard layout for that. 2010-09-27T14:19:27 colemak might be a bit better for speed 2010-09-27T14:19:37 since it emphasizes "finger rolls" and stuff 2010-09-27T14:20:23 better to switch before to have an health reason to do it. 2010-09-27T14:20:37 it also preserves some of your keyboard shortcuts from qwerty, but i don't know if i care much about that 2010-09-27T14:20:50 Thanks for the info 2010-09-27T14:21:07 elijahbal: exactly my thinking 2010-09-27T14:21:15 although i still haven't made the switch 2010-09-27T14:21:18 ah yes, that was the reason I chose programmer dvorak over regular dvorak, the x c and v keys are in the same spot 2010-09-27T14:21:41 jmcarthur: Colemak is designed with finger rolls in mind? I've never heard that. 2010-09-27T14:21:41 no wait they aren't XD 2010-09-27T14:21:53 mightybyte: at least in part 2010-09-27T14:22:15 All I've seen is preserving zxcv and symbols. 2010-09-27T14:22:23 ah yes, it's like this: while I hold control, it switches back to qwerty 2010-09-27T14:22:28 im going for dvorak 2010-09-27T14:22:38 *** tei has joined #aichallenge 2010-09-27T14:22:53 jmcarthur: if you are lazy like me, you'll probably want to wait until your fingers get broken. 2010-09-27T14:22:54 than i can choose between vim/emacs/others (: 2010-09-27T14:23:15 tired of cumbersome ides 2010-09-27T14:23:33 HakanD: I found that the easiest way to use vim with dvorak was to not do any key remapping. 2010-09-27T14:23:40 else vim is designed to make your own shortcuts. 2010-09-27T14:23:59 mightybyte: sure ? 2010-09-27T14:24:05 elijahbal: For me, yes. 2010-09-27T14:24:19 l and h are pretty intuitive for left and right movement. 2010-09-27T14:24:34 (although slightly less convenient than their qwerty positions) 2010-09-27T14:24:57 It wasn't too hard to re-learn jk 2010-09-27T14:25:31 Although I must say that even though I can still type pretty well on qwerty, using vim in qwerty is tough. 2010-09-27T14:26:49 vim for c++ / python ? 2010-09-27T14:26:51 *** Itkovian has joined #aichallenge 2010-09-27T14:27:37 what do you prefer for java? 2010-09-27T14:30:25 *** amstan has joined #aichallenge 2010-09-27T14:30:25 *** ChanServ sets mode: +o amstan 2010-09-27T14:30:27 HakanD: I use emacs and Dvorak 2010-09-27T14:30:57 it's great except for copying and pasting, as was mentioned earlier 2010-09-27T14:31:32 I did it to reduce the RSI (repetitive stress injure) I was beginning to develop 2010-09-27T14:31:47 and it worked. The other main thing though was capswap 2010-09-27T14:32:12 k, switched to dvorak, and typing this took around 2 3 mins :) 2010-09-27T14:32:17 L-Ctrl and Caps Lock are swapped 2010-09-27T14:32:21 ha right 2010-09-27T14:32:28 it's slow at first but I now type 85 wpm 2010-09-27T14:32:53 it took a few months to match my QWERTY speed 2010-09-27T14:33:07 HakanD: You should start with some tutorial which introduces the keys incrementally to let you build up familiarity with them. 2010-09-27T14:33:24 yep im gonna to that 2010-09-27T14:33:25 setxkbmap dvorak option ctrl:swapcaps 2010-09-27T14:33:30 just wanted to try (: 2010-09-27T14:33:44 HakanD: How fast do you type in qwerty? 2010-09-27T14:33:47 that enables dvorak and swaps l-ctrl with caps lock in Linux 2010-09-27T14:34:07 90-100 tops 2010-09-27T14:34:14 but that was 4-5 months ago 2010-09-27T14:34:21 im a bit slow right now 2010-09-27T14:34:30 didnt typed a lot (: 2010-09-27T14:35:12 after reading a lot about rsi, i decided to take action early 2010-09-27T14:35:25 HakanD: Ok. I would guess that you should be able to get 40 or so wpm on dvorak with a few hours effort (or less). 2010-09-27T14:35:41 *** bhasker_ has joined #aichallenge 2010-09-27T14:36:00 Thanks for great info again 2010-09-27T14:36:09 Sure 2010-09-27T14:36:28 HakanD: when switching to Dvorak, debating people over IM really motivated me to type fast, ha 2010-09-27T14:36:38 (: 2010-09-27T14:36:39 timed typing tests are good too 2010-09-27T14:37:15 anyone have some idea how to debug scripts on PHP on TCP server ? 2010-09-27T14:37:21 *** bhasker has quit IRC (Ping timeout: 252 seconds) 2010-09-27T14:37:54 u1tr0n: stderr? a log file? 2010-09-27T14:38:07 *** elimisteve has left #aichallenge ("Leaving") 2010-09-27T14:39:25 MyBot vs DualBot 100/100 Avg Turns: 60.0, MyBot vs RageBot 98/100 Avg Turns: 54.0 2010-09-27T14:39:57 how can you lose to RageBot after you block its first assault? 2010-09-27T14:40:39 well, if you don't block its first assault... 2010-09-27T14:40:56 oh ok 2010-09-27T14:41:16 *** bhasker has joined #aichallenge 2010-09-27T14:41:25 impressive that you still got 98/100 then, as there are quite a few maps with close starting planets 2010-09-27T14:41:39 how do you calculate these numbers? 2010-09-27T14:42:29 one of the 2 is from being 2 ships short of blocking and the other is from being to greedy and not saving ships to block the second one (map 22 43 turns and map 30 200 turns) 2010-09-27T14:42:38 a script from greghaynes 2010-09-27T14:42:40 *** justin_pdx has quit IRC (Quit: justin_pdx) 2010-09-27T14:42:54 stress testing scripts against all maps 2010-09-27T14:43:02 before every upload 2010-09-27T14:43:04 *** justin_pdx has joined #aichallenge 2010-09-27T14:43:25 http://ai-contest.com/visualizer.php?game_id=4849694 2010-09-27T14:43:33 *** bhasker_ has quit IRC (Ping timeout: 240 seconds) 2010-09-27T14:44:06 *** iouri_ has joined #aichallenge 2010-09-27T14:45:07 the top people must be doing 100/100 and working on lowering turns or something 2010-09-27T14:45:37 *** tapwater has joined #aichallenge 2010-09-27T14:52:24 i don't currently have a top bot, but i don't plan on focusing on lowering turns unless the game objective changes somewhere down the line 2010-09-27T14:53:17 that is, a low turn count for a win may be a good sign, but it's not really the goal 2010-09-27T14:54:17 *** batman0730 has joined #aichallenge 2010-09-27T14:56:33 *** HakanD has quit IRC (Ping timeout: 240 seconds) 2010-09-27T14:57:46 Top 10 players: dmj111(3683), adumlah(3527), delineate(3494), sequoh(3473), bix0r4ever(3469), kcm1700(3451), davidjliu(3440), sighbot(3417), McLeopold(3407), Rexxar(3405) 2010-09-27T14:58:30 well, they're probably working on improving against other good bots 2010-09-27T14:58:47 at a certain level, the example bots become irrelevant 2010-09-27T15:01:20 dmj111 is very much nicer 2010-09-27T15:09:38 *** batman0730 has quit IRC (Ping timeout: 252 seconds) 2010-09-27T15:10:54 *** jens has joined #aichallenge 2010-09-27T15:11:28 *** jens is now known as Guest56168 2010-09-27T15:13:32 last year top 10 places were C++ bots 2010-09-27T15:13:37 any prediction this year? :) 2010-09-27T15:14:49 *** fahi has left #aichallenge 2010-09-27T15:15:19 *** bhasker has quit IRC (Quit: bhasker) 2010-09-27T15:16:02 tei: so far as I've seen, AI for PW doesn't require as raw computing power as Tron; my bot still doesn't use it's allotted 1sec/move. I'm guessing there's a lot more room for higher languages here. 2010-09-27T15:20:40 *** bhasker has joined #aichallenge 2010-09-27T15:20:44 its not a solvable problem like tron became 2010-09-27T15:21:35 *** Accoun has quit IRC () 2010-09-27T15:25:29 *** Cyndre_ has joined #aichallenge 2010-09-27T15:27:26 it doesn't seem like anyone has come up with a viable scalable algorithm for dealing with planetwars yet 2010-09-27T15:27:37 working on it :P 2010-09-27T15:27:44 I have my ideas 2010-09-27T15:28:09 great, I really hope that the winner ends up using one 2010-09-27T15:29:04 *** Cyndre has quit IRC (Ping timeout: 252 seconds) 2010-09-27T15:29:22 yeah, as there are some rock-paper-scissors mechanics in this game it'd suck if the winner just won because he had good parameters and some luck 2010-09-27T15:30:25 *** amstan has quit IRC (Ping timeout: 265 seconds) 2010-09-27T15:32:57 but then again, perhaps there were rock-paper-scissors mechanics in tron, I haven't thought about that as much 2010-09-27T15:33:18 *** TapewormGalactos has joined #aichallenge 2010-09-27T15:35:04 there may have been some r-p-s situations depending on the map but they were completely swamped out by the need to have very good search 2010-09-27T15:38:08 actually I should say, there were r-p-s situations... 2010-09-27T15:38:32 *** Blkt has quit IRC (Ping timeout: 276 seconds) 2010-09-27T15:40:34 *** RainCT has joined #aichallenge 2010-09-27T15:43:41 *** boegel has joined #aichallenge 2010-09-27T15:45:13 *** bhasker_ has joined #aichallenge 2010-09-27T15:47:12 *** Accoun has joined #aichallenge 2010-09-27T15:47:55 if only move iteration wasn't so hard I'd have viable scalable algorithm already... 2010-09-27T15:48:40 ? 2010-09-27T15:48:59 *** bhasker has quit IRC (Ping timeout: 255 seconds) 2010-09-27T15:49:20 nvm 2010-09-27T15:50:33 *** Guest56168 has quit IRC (Ping timeout: 240 seconds) 2010-09-27T15:50:38 http://ai-contest.com/profile.php?user_id=3911 ooooo looking good so far 2010-09-27T15:55:54 *** roku151 has joined #aichallenge 2010-09-27T15:57:46 Top 10 players: dmj111(3636), adumlah(3538), delineate(3510), sequoh(3486), bix0r4ever(3479), kcm1700(3472), Rexxar(3464), davidjliu(3452), sighbot(3436), McLeopold(3425) 2010-09-27T16:00:57 *** delt0r___ has joined #aichallenge 2010-09-27T16:01:36 *** jens has joined #aichallenge 2010-09-27T16:02:02 *** jens is now known as Guest27250 2010-09-27T16:02:39 *** Naith has quit IRC (Quit: Naith) 2010-09-27T16:02:53 rank 123 with a bot that lost a few games due to time-outs 2010-09-27T16:02:57 this is promising... 2010-09-27T16:03:05 *** delt0r_ has quit IRC (Ping timeout: 272 seconds) 2010-09-27T16:09:14 grats 2010-09-27T16:09:47 boegel: try it on the zeus server please :) 2010-09-27T16:09:57 *** Guest27250 is now known as JensTi 2010-09-27T16:10:38 on the zeus server? 2010-09-27T16:10:53 *** tei has quit IRC (Ping timeout: 252 seconds) 2010-09-27T16:12:01 *** Naktibalda has joined #aichallenge 2010-09-27T16:13:47 *** aerique has joined #aichallenge 2010-09-27T16:20:51 *** Bobng has quit IRC (Quit: Leaving) 2010-09-27T16:25:40 *** amstan has joined #aichallenge 2010-09-27T16:25:41 *** ChanServ sets mode: +o amstan 2010-09-27T16:26:08 *** JensTi has quit IRC (Quit: Leaving.) 2010-09-27T16:28:48 *** amstan has quit IRC (Remote host closed the connection) 2010-09-27T16:31:08 *** timintal has joined #aichallenge 2010-09-27T16:33:14 *** BtbN has joined #aichallenge 2010-09-27T16:38:15 *** roku151 has quit IRC (Quit: Page closed) 2010-09-27T16:38:15 *** megalomania has joined #aichallenge 2010-09-27T16:43:21 *** Naith has joined #aichallenge 2010-09-27T16:45:03 *** timintal has quit IRC (Quit: Page closed) 2010-09-27T16:51:17 *** sinan has joined #aichallenge 2010-09-27T16:57:47 Top 10 players: dmj111(3630), adumlah(3532), delineate(3505), sequoh(3482), bix0r4ever(3472), kcm1700(3471), Rexxar(3456), davidjliu(3445), sighbot(3431), felixcoto(3425) 2010-09-27T16:59:04 *** Olathe has joined #aichallenge 2010-09-27T16:59:37 hmm 2010-09-27T16:59:41 *** JensTi has joined #aichallenge 2010-09-27T16:59:59 I'm seem to be losing multiple games due to time-outs (don't think it's invalid moves, I checked that) 2010-09-27T17:00:45 but, when I replay the game using parse_game_state_url.py, I see the bot process all the game states in less than 50ms 2010-09-27T17:01:14 while a time-out should only occur when I spend more than 1s in some turn, right? 2010-09-27T17:01:26 boegel: you can also time out if you send bad commands 2010-09-27T17:01:31 is the server still susceptible to 'freak' time out? 2010-09-27T17:01:33 like sending from planets that do not belong to you 2010-09-27T17:01:44 JensTi: yes, I know, I'm pretty sure I don't do that 2010-09-27T17:01:46 or sending more ships then there are on a planet 2010-09-27T17:02:32 JensTi: yes, I'm aware of all that, but my best guess now is a freak server time-out, to be honest 2010-09-27T17:03:06 boegel: if the time out is on the 1st turn, then yes 2010-09-27T17:03:50 JensTi: even then 2010-09-27T17:04:16 JensTi: the server could become very busy doing other stuff during any given turn, which would screw up the time-out mechanism imho 2010-09-27T17:04:26 JensTi: not sure how many measures have been taken against that 2010-09-27T17:05:30 example of such a game: http://ai-contest.com/visualizer.php?game_id=4855987 2010-09-27T17:05:47 can some admin check whether this was a time-out or an invalid move? 2010-09-27T17:06:12 *** Palmik has quit IRC (Remote host closed the connection) 2010-09-27T17:08:11 *** Mathnerd314 has joined #aichallenge 2010-09-27T17:13:54 *** elijahbal has quit IRC (Ping timeout: 240 seconds) 2010-09-27T17:14:39 *** elijahbal has joined #aichallenge 2010-09-27T17:15:00 *** justin_pdx has quit IRC (Quit: justin_pdx) 2010-09-27T17:15:34 *** smellyhippy has quit IRC (Ping timeout: 240 seconds) 2010-09-27T17:19:26 *** anubhav_ has joined #aichallenge 2010-09-27T17:19:41 *** smellyhippy has joined #aichallenge 2010-09-27T17:19:41 *** smellyhippy has joined #aichallenge 2010-09-27T17:20:01 *** amstan has joined #aichallenge 2010-09-27T17:20:01 *** ChanServ sets mode: +o amstan 2010-09-27T17:27:52 *** sinan has quit IRC (Ping timeout: 252 seconds) 2010-09-27T17:29:52 *** megalomania has quit IRC (Quit: Leaving.) 2010-09-27T17:32:11 *** rogue780 has quit IRC (Quit: Leaving) 2010-09-27T17:34:20 *** sigh has joined #aichallenge 2010-09-27T17:37:52 *** davidjliu has joined #aichallenge 2010-09-27T17:42:00 *** Appleman1234 has joined #aichallenge 2010-09-27T17:49:09 *** mega1 has quit IRC (Ping timeout: 255 seconds) 2010-09-27T17:50:13 amstan: are you aware of any issues with bots timing out on the official server, while they really shouldn't be? 2010-09-27T17:50:24 nope, sorry 2010-09-27T17:51:08 hmm, crap 2010-09-27T17:51:49 amstan: can you check why I lost this particular game: http://ai-contest.com/visualizer.php?game_id=4832591 ? 2010-09-27T17:51:56 boegel: i have no idea how 2010-09-27T17:52:02 amstan: is it indeed a time-out of my bot, or something else? 2010-09-27T17:52:08 amstan: ah, you can' check? 2010-09-27T17:52:24 s/can'/can't/ 2010-09-27T17:52:31 I assumed you're one of the admins :) 2010-09-27T17:53:12 boegel: unless someone changed it recently I don't think the server records how a game ended 2010-09-27T17:53:23 Janzert: ah :( 2010-09-27T17:53:26 boegel: i am, but idk how 2010-09-27T17:54:25 I just find it really, really strange, because the exact same bot playing at the TCP server doesn't time out at all 2010-09-27T17:55:22 If the server is under heavy load, a simple calculation might take forever and cause a timeout 2010-09-27T17:55:25 have you tried testing locally with the same version of haskell as is on the contest server? 2010-09-27T17:55:33 it realy should check for CPU-Time, and not real time 2010-09-27T17:55:34 maybe it's a version difference? 2010-09-27T17:56:08 CPU time has it's own problems too 2010-09-27T17:56:17 not as vast as wall clock time has though 2010-09-27T17:56:27 Janzert: I was thinking the exact same thing now 2010-09-27T17:57:20 personally I think ideal is wall clock time on servers dedicated to only running a game at a time 2010-09-27T17:57:36 *** aerique has quit IRC (Quit: ...) 2010-09-27T17:57:46 Top 10 players: dmj111(3581), adumlah(3493), Rexxar(3460), sighbot(3453), davidjliu(3443), Loks1982(3435), sequoh(3422), felixcoto(3410), kcm1700(3398), McLeopold(3370) 2010-09-27T17:58:08 *** iouri_ has quit IRC (Quit: Page closed) 2010-09-27T17:58:16 Janzert: as long as each bot gets a core, i agree 2010-09-27T17:58:18 actually ideal is dedicated server per player 2010-09-27T17:58:24 yeah 2010-09-27T17:59:06 with identical specs, of course 2010-09-27T17:59:06 but then you also need to make sure servers are matched and both on same local network as engine server 2010-09-27T17:59:38 and of course each step is way more resource intensive 2010-09-27T17:59:47 :( 2010-09-27T17:59:55 something like the tcp server but with an army of dedicated, local bot servers rather than individual home machines 2010-09-27T18:00:05 * Janzert nods 2010-09-27T18:00:16 each bot server would just cycle through a list and connect to the main server for matches 2010-09-27T18:02:01 1 server per bot, right? 2010-09-27T18:02:19 we all want more games 2010-09-27T18:02:26 well 1 bot per server at least 2010-09-27T18:02:29 at a time 2010-09-27T18:03:52 someone needs to get Dr. Robert Hyatt the author of the chess program Crafty interested in the contest. 2010-09-27T18:04:17 I know he uses a cluster for testing crafty that he can do 30000 full chess games in a few hours 2010-09-27T18:04:46 Janzert: I have that for this comp... I have a 512 core cluster... 2010-09-27T18:04:50 I still suck 2010-09-27T18:05:00 hehe 2010-09-27T18:05:03 :'( 2010-09-27T18:05:30 * boegel has access to a 1000+ core server, but is unsure if he's allowed to used it for this stuff 2010-09-27T18:05:59 * Zannick has access to... 2010-09-27T18:06:01 >.> 2010-09-27T18:06:03 I am as long as i put it in the low priority que 2010-09-27T18:06:52 be nice if someone could get access to run the next contest matches on one 2010-09-27T18:07:08 <.< 2010-09-27T18:07:36 I have a P90 sitting here to run tests on :) 2010-09-27T18:07:44 cool 2010-09-27T18:07:50 My 486 finally dies 2010-09-27T18:07:54 died 2010-09-27T18:08:28 It was a Doom II monster --with 8Meg of ram 2010-09-27T18:08:50 heh 2010-09-27T18:08:51 cost me over 150EU to upgrade it from 4Meg 2010-09-27T18:10:18 *** RainCT has quit IRC (Remote host closed the connection) 2010-09-27T18:10:26 Euros didn't exist in these days 2010-09-27T18:12:25 I converted 2010-09-27T18:12:31 I was in NZ at the time 2010-09-27T18:12:32 *** justin_pdx has joined #aichallenge 2010-09-27T18:12:51 it was in fact about 350NZD 2010-09-27T18:13:51 *** Arthur_ has quit IRC (Ping timeout: 276 seconds) 2010-09-27T18:16:20 *** Arthur_ has joined #aichallenge 2010-09-27T18:18:04 confused how is it that on certain maps my bot can beat my own bot only in one position 2010-09-27T18:18:17 i would expect my bot to draw everytime 2010-09-27T18:18:31 bug or hidden randomness 2010-09-27T18:18:47 ie some order is important somewhere (like say fleets) 2010-09-27T18:18:54 Thats my guess 2010-09-27T18:19:10 hmm 2010-09-27T18:19:21 also planet order is given the same for both players 2010-09-27T18:19:41 Or you go through planets left to right... and that affects behavour ie how ties are broken 2010-09-27T18:19:52 *** nrub has quit IRC (Read error: Connection reset by peer) 2010-09-27T18:20:14 for example 2010-09-27T18:21:08 my bot also beats himself, but i don't think that's unusual 2010-09-27T18:21:16 I have some randomness by the fact that i iterate through a hash map 2010-09-27T18:21:31 one has the better starting-position for the strategy they use 2010-09-27T18:21:39 but player 1 and 2 would be roughly even however 2010-09-27T18:22:10 BtbN: the maps have rotational symmetry 2010-09-27T18:22:39 all? 2010-09-27T18:22:52 yes 2010-09-27T18:22:54 for me they looked quite random o_O 2010-09-27T18:22:56 so far 2010-09-27T18:23:16 my bot allways managed to beat himself 2010-09-27T18:23:24 in all of his versions 2010-09-27T18:23:30 and i'm not using any random stuff 2010-09-27T18:23:43 Each planet is placed --then 180deg around the center another identical planet is placed 2010-09-27T18:24:10 + one center planet? 2010-09-27T18:24:12 however that could lead to different rounding behavour for distances 2010-09-27T18:24:23 BtbN: yea --sometimes 2010-09-27T18:24:30 didn't check that 2010-09-27T18:24:35 i thought all maps have 23 planets? 2010-09-27T18:24:37 *** jaspervdj has quit IRC (Quit: NEVER GONNA GIVE YOU UP NEVER GONNA LET YOU DOWN) 2010-09-27T18:24:43 I guess 2010-09-27T18:24:45 theres always a center 2010-09-27T18:24:54 I don't hard code anything... 2010-09-27T18:24:56 BtbN: yes, it's true 2010-09-27T18:25:18 hm, wtf does it beat itself then o_O 2010-09-27T18:25:34 it even did so in it's simple version 2010-09-27T18:25:43 watch a game and find differences :) 2010-09-27T18:26:00 capture of central planet breaks symetry 2010-09-27T18:26:05 they do a complete differnnt beginning 2010-09-27T18:26:09 capture other planets 2010-09-27T18:26:10 *** boegel has quit IRC (Quit: Leaving) 2010-09-27T18:26:22 the game itself does not look symetrical 2010-09-27T18:26:26 then your bot is unpredictable :) 2010-09-27T18:26:52 can you repeat a game? is result the same? 2010-09-27T18:27:08 yep, all games are fully repeatable 2010-09-27T18:27:26 record game to file and upload it so we can see it 2010-09-27T18:27:29 wasn't there a "game-pastbin" somewhere? 2010-09-27T18:27:35 http://visualizer.naktibalda.lt/ 2010-09-27T18:27:36 I don't think fleet orders are the same between players... 2010-09-27T18:27:54 ie player 1 is always first... but thats player 2 for player 2... 2010-09-27T18:28:15 Naktibalda: does it understand the log.txt of the cpp-visualiser? 2010-09-27T18:28:23 hmm 2010-09-27T18:28:29 redirect to file instead of visualizer 2010-09-27T18:28:40 java -jar playGame .... >file.txt 2010-09-27T18:28:42 the visualizer is all-in-one 2010-09-27T18:29:14 BtbN: there are rounding errors in the distance 2010-09-27T18:29:17 if it's the same as log.txt of playGame then no 2010-09-27T18:29:29 nah, it does not take it 2010-09-27T18:29:31 wait a second 2010-09-27T18:29:42 stupid windows cmd.exe 2010-09-27T18:29:47 need bash 2010-09-27T18:30:00 JensTi: in map or in bot code? 2010-09-27T18:31:58 in bot code 2010-09-27T18:32:10 Naktibalda: http://visualizer.naktibalda.lt/71 2010-09-27T18:32:19 it's symetric at the beginning 2010-09-27T18:34:07 BtbN: that's because you do a for loop over a collection somewhere ? 2010-09-27T18:34:45 yeah, a lot of them 2010-09-27T18:34:51 foreach-loops 2010-09-27T18:34:55 but no random maps or so 2010-09-27T18:35:09 if you somewhere select the 1st entry in the list and ignore the rest it may give some random effects 2010-09-27T18:35:26 only in the list of owned planets 2010-09-27T18:36:16 because if the list is sorted in some way, so that the first and second planet have the same amount of troops f.ex 2010-09-27T18:36:31 you might end up with first planet a in bot 1 and planet b in bot 2 2010-09-27T18:36:50 because memory allocation is kind of random 2010-09-27T18:36:52 i can't realy find the first move they desync 2010-09-27T18:36:53 *** Migi32 has quit IRC (Quit: http://boinc.berkeley.edu) 2010-09-27T18:37:48 BtbN: they both send troops to the same planet, with 81 troops on the blue side 2010-09-27T18:38:06 *** Naith has quit IRC (Quit: Naith) 2010-09-27T18:38:13 because it is closer to blue he gets there first and has spawned more fighters before red arives 2010-09-27T18:38:18 so there they desync 2010-09-27T18:38:53 so an optimalisation might be to check the distance to the planet you send to, and first send to the closest ;) 2010-09-27T18:38:54 hm, so it's because they both pick the planet with the lower ID, and it's closer to bot 1 2010-09-27T18:39:00 yep 2010-09-27T18:39:13 i should sort the list by distance to home-plane 2010-09-27T18:39:25 *** wh1teside has joined #aichallenge 2010-09-27T18:39:33 uhu, in theory 2010-09-27T18:39:50 but in practice you will be faster just doing a check somewhere in the for loop 2010-09-27T18:39:55 instead of sorting the whole thing 2010-09-27T18:40:41 do you sort on troops? or do a loop over all? 2010-09-27T18:40:42 it's quite stupid, it goes through the NotMine-Planets, and attacks the first one he has enough ships for to attack 2010-09-27T18:41:09 BtbN: You should combine forces from several planets 2010-09-27T18:41:10 and this list is sorted by id 2010-09-27T18:41:30 Itkovian: if the bot would work as intended there wouldn't be any forces outside the base-planet :D 2010-09-27T18:41:34 BtbN: the way i work is by computing some sort of score for every notmine planets 2010-09-27T18:41:40 did not yet figure out why it sometimes forgetts some 2010-09-27T18:41:50 based on number of fleets, then distance, then wether neutral or owned by other 2010-09-27T18:42:06 and pick the biggest in there 2010-09-27T18:42:35 yeah, i thought about something like that already 2010-09-27T18:44:56 yeah, i see the place in my code where it fails 2010-09-27T18:45:06 some scoring-algorithm should fix it 2010-09-27T18:48:46 yep, they draw now 2010-09-27T18:48:47 nice 2010-09-27T18:49:47 wow, and it also beats RageBot now 2010-09-27T18:52:44 Naktibalda: I wrote a bot that did things randomly once. It's possible to have non-repeatable matches. But you can always view a specific games output. 2010-09-27T18:53:13 I know that it's possible, but that's not very common 2010-09-27T18:53:27 For good reason. Ask me how well that bot did. 2010-09-27T18:54:27 choosing random target from top 3 targets has no advantages against choosing top target (except unpredictability) 2010-09-27T18:57:47 Top 10 players: dmj111(3622), adumlah(3504), Rexxar(3488), davidjliu(3457), sequoh(3455), felixcoto(3453), sighbot(3450), Loks1982(3450), kcm1700(3416), McLeopold(3382) 2010-09-27T18:59:43 ah, compile-mails work again 2010-09-27T19:02:56 *** sigh has joined #aichallenge 2010-09-27T19:03:12 *** anoek has joined #aichallenge 2010-09-27T19:10:46 *** boiserm has joined #aichallenge 2010-09-27T19:11:25 *** bhasker has joined #aichallenge 2010-09-27T19:12:54 a1k0n: I just read your postmortem from last year's contest. Impressive! Thanks for taking the time to write all of that out 2010-09-27T19:13:44 *** bhasker_ has quit IRC (Ping timeout: 272 seconds) 2010-09-27T19:14:25 *** bhasker_ has joined #aichallenge 2010-09-27T19:17:19 *** bhasker has quit IRC (Ping timeout: 245 seconds) 2010-09-27T19:24:15 *** medice has quit IRC (Ping timeout: 276 seconds) 2010-09-27T19:24:19 *** medice has joined #aichallenge 2010-09-27T19:31:29 *** elijahbal has quit IRC (*.net *.split) 2010-09-27T19:31:29 *** Naktibalda has quit IRC (*.net *.split) 2010-09-27T19:31:29 *** jesusaurus has quit IRC (*.net *.split) 2010-09-27T19:31:29 *** astebbin has quit IRC (*.net *.split) 2010-09-27T19:35:40 *** elijahbal has joined #aichallenge 2010-09-27T19:35:40 *** Naktibalda has joined #aichallenge 2010-09-27T19:35:40 *** jesusaurus has joined #aichallenge 2010-09-27T19:35:40 *** astebbin has joined #aichallenge 2010-09-27T19:40:31 boiserm: glad you liked it! 2010-09-27T19:43:10 *** JensTi has quit IRC (Quit: Leaving.) 2010-09-27T19:50:06 *** jimmerton has joined #aichallenge 2010-09-27T19:50:23 quick questions, anyone with a macbook pro 2010-09-27T19:50:29 *question 2010-09-27T19:50:45 do you have 5 grey bars beneath the screen on the left hand side? 2010-09-27T19:50:55 just below the visible area of the monitor 2010-09-27T19:51:07 uh 2010-09-27T19:51:28 i don't actually have a macbook pro but i used one at work and don't recall what you are talking about 2010-09-27T19:52:16 I have one but I don't know what you're talking about. 2010-09-27T19:53:01 weird, it's just outside of the visible area 2010-09-27T19:53:05 it's odd 2010-09-27T19:53:12 I guess it's a manufacturing defect 2010-09-27T19:53:21 it's not really a bother 2010-09-27T19:57:47 Top 10 players: dmj111(3688), adumlah(3506), Rexxar(3490), davidjliu(3488), sighbot(3470), sequoh(3463), Loks1982(3413), kcm1700(3402), vinx33(3401), wagstaff(3392) 2010-09-27T19:59:09 a1k0n: I'd like to thank you too, I wasn't in the tron contest, but that was an interesting read :) 2010-09-27T20:01:34 Itkovian!! - A steven erikson fan 2010-09-27T20:04:20 Definitely 2010-09-27T20:04:26 But one who is due for sleep 2010-09-27T20:04:29 overdue even 2010-09-27T20:05:33 heh 2010-09-27T20:05:37 i can't wait for the 10th book 2010-09-27T20:05:52 and esselemont's next book 2010-09-27T20:05:56 *** BtbN has quit IRC (Quit: Verlassend) 2010-09-27T20:10:44 :( Eclipse autocomplete just suddenly stopped working 2010-09-27T20:11:27 restarting Eclipse didn't help either... back to Vim I guess 2010-09-27T20:20:01 ...Fleet is not defined? 2010-09-27T20:20:02 *** ichti has joined #aichallenge 2010-09-27T20:20:06 :( 2010-09-27T20:20:14 sigh: not you, just a general sigh... 2010-09-27T20:20:18 :P 2010-09-27T20:20:32 :P 2010-09-27T20:20:34 o.O 2010-09-27T20:20:42 @sigh 2010-09-27T20:20:42 i know... 2010-09-27T20:21:22 is there a difference bewteen the online game engine, and the one in the starter pack (or the 1.2 I got from the forum) 2010-09-27T20:21:34 ichti: online they use the python engine 2010-09-27T20:22:40 ok, but my bot lost oline on 1st turn somehow, but it doesn't when I run it my self. 2010-09-27T20:22:45 I don't like that :( 2010-09-27T20:22:57 it means you timed out or made an illegal move 2010-09-27T20:23:03 *** sigh has quit IRC (Remote host closed the connection) 2010-09-27T20:23:17 *** sigh has joined #aichallenge 2010-09-27T20:23:31 I know, but I can't make that happen when I run it at my pc. 2010-09-27T20:23:57 ichti: maybe you're using a different version of your programming language than the server? 2010-09-27T20:24:27 I'm using python27, what's the server using? :P 2010-09-27T20:24:34 2.5 I think 2010-09-27T20:24:34 2.5 2010-09-27T20:24:38 *** jesionaj` is now known as jesionaj 2010-09-27T20:24:41 meh 2010-09-27T20:25:28 @help 2010-09-27T20:25:28 boiserm: (help [] []) -- This command gives a useful description of what does. is only necessary if the command is in more than one plugin. 2010-09-27T20:25:58 !rainbow contestbot help is not that helpful 2010-09-27T20:25:58 sigh: contestbot help is not that helpful 2010-09-27T20:26:03 @makeMeAWinningBot 2010-09-27T20:26:03 boiserm: An error occured while trying to show the previous error. 2010-09-27T20:26:51 contestbot: rankings 100000 2010-09-27T20:26:51 sigh: Error: Cannot do that many players. 2010-09-27T20:27:01 contestbot: rankings 50 2010-09-27T20:27:01 sigh: Error: Cannot do that many players. 2010-09-27T20:27:05 contestbot: rankings 49 2010-09-27T20:27:07 sigh: Top 49 players: dmj111(3682), adumlah(3507), davidjliu(3483), Rexxar(3480), sighbot(3466), sequoh(3458), Loks1982(3404), vinx33(3401), kcm1700(3400), McLeopold(3387), wagstaff(3374), delineate(3357), user42(3340), voidptr(3329), drewrobb(3316), ImpBot(3314), montanalow(3314), r.nitsch(3309), DocBot(3308), extrapostrophe(3302), ClusterTeam(3298), GreenTea(3295), smatcher(3290), paulus(3287), (1 more message) 2010-09-27T20:27:21 what an odd cuttoff 2010-09-27T20:27:41 contestbot: rankings 0 2010-09-27T20:27:41 justin_pdx: Error: Cannot do that many players. 2010-09-27T20:27:56 contestbot: how many players can you do? 2010-09-27T20:27:57 sigh: Run as fast as you can and don't look back. 2010-09-27T20:28:05 oddly appropriate 2010-09-27T20:29:04 lol 2010-09-27T20:33:19 contestbot: commands 2010-09-27T20:33:19 boiserm: action, add, alert, alias, allyourspace, announce, announce add, announce list, announce remove, any, aol, apply, apropos, at, author, azn, ban add, ban list, ban remove, base, binary, bold, bye, cache, calc, call, capabilities, capability add, capability list, capability remove, capability set, capability setdefault, capability unset, capitalize, change, changename, channel, channels, (5 more messages) 2010-09-27T20:33:24 hey, gold mine 2010-09-27T20:33:33 first try 2010-09-27T20:33:39 err... second try :) 2010-09-27T20:34:12 contestbot: capability list 2010-09-27T20:34:12 boiserm: -halfop -op -protected -voice 2010-09-27T20:34:21 not what I expected 2010-09-27T20:34:35 eh, whatever 2010-09-27T20:35:28 contestbot: rankings -2 2010-09-27T20:35:28 Cyndre_: Error: Cannot do that many players. 2010-09-27T20:35:43 contestbot: rankings 0x000001 2010-09-27T20:35:45 Cyndre_: Top 1 players: dmj111(3683) 2010-09-27T20:35:54 thats probably not a good thing 2010-09-27T20:36:25 contestbot: rankings 0x00000f 2010-09-27T20:36:26 Cyndre_: Top 15 players: dmj111(3683), adumlah(3511), davidjliu(3483), sighbot(3464), sequoh(3460), Rexxar(3438), vinx33(3427), kcm1700(3403), Loks1982(3385), McLeopold(3385), wagstaff(3373), delineate(3361), user42(3337), voidptr(3333), DocBot(3314) 2010-09-27T20:36:38 :) 2010-09-27T20:37:08 @list aichallenge 2010-09-27T20:37:08 sigh: game, match, and rankings 2010-09-27T20:38:52 now the real question is "Is contestbot sanitary?" 2010-09-27T20:39:06 contestbot: Are you sanitary? 2010-09-27T20:39:07 justin_pdx: User error -- Replace user. 2010-09-27T20:39:35 if you accept hex inplace of integer what are the chances that your sanitary? 2010-09-27T20:40:53 it working for hex doesnt mean too much 2010-09-27T20:41:19 most languages (and presumably sql) accept multiple formats for numbers 2010-09-27T20:41:45 should really be hard coded - only integers 2010-09-27T20:41:52 @rankings 0b111 2010-09-27T20:41:53 sigh: Top 7 players: dmj111(3685), adumlah(3513), davidjliu(3484), sighbot(3465), sequoh(3461), Rexxar(3438), vinx33(3428) 2010-09-27T20:42:05 but 0x0f is an integer 2010-09-27T20:42:17 jesionaj: would fail a strict check 2010-09-27T20:42:36 @rankings 015 2010-09-27T20:42:37 sigh: Top 13 players: dmj111(3683), adumlah(3510), davidjliu(3482), sighbot(3463), sequoh(3459), Rexxar(3435), vinx33(3426), kcm1700(3403), McLeopold(3384), Loks1982(3382), delineate(3361), wagstaff(3353), user42(3338) 2010-09-27T20:42:39 this is fun :) 2010-09-27T20:42:49 015 = 13 :D 2010-09-27T20:42:59 you sure about that? 2010-09-27T20:43:05 wait for ranking mysql injection here 2010-09-27T20:43:06 lol 2010-09-27T20:43:14 nah, the code is safe 2010-09-27T20:44:01 it just does an int() on the input string, i think 2010-09-27T20:44:05 http://ai-contest.com/visualizer.php?game_id=4864274 how did I blow this 2010-09-27T20:44:22 @ranking 12kd 2010-09-27T20:44:22 Cyndre_: (rankings [how many]) -- Get the top $(how many) players. 2010-09-27T20:44:26 lmao 2010-09-27T20:44:45 it doesnt like me 2010-09-27T20:46:24 *** ichti has quit IRC (Quit: Page closed) 2010-09-27T20:47:29 @sigh 2010-09-27T20:47:29 i know... 2010-09-27T20:47:33 @sigh @sigh 2010-09-27T20:47:33 i know... 2010-09-27T20:47:37 :-P 2010-09-27T20:47:45 ... 2010-09-27T20:48:01 contestbot: help sigh 2010-09-27T20:48:01 sigh: (sigh ) -- Alias for "echo i know...". 2010-09-27T20:48:11 lol 2010-09-27T20:48:36 action 2010-09-27T20:48:44 @contest 2010-09-27T20:48:45 Cyndre_: An error occured while trying to show the previous error. 2010-09-27T20:48:50 lmao 2010-09-27T20:48:59 @aol 2010-09-27T20:48:59 Cyndre_: (aol ) -- Returns as if an AOLuser had said it. 2010-09-27T20:49:05 haha 2010-09-27T20:49:19 !seen Cyndre_ 2010-09-27T20:49:19 sigh: Cyndre_ was last seen in #aichallenge 19 seconds ago: @aol 2010-09-27T20:49:22 @aol this is a test... a test of the emergency AOLuser system 2010-09-27T20:49:22 boiserm: this is a test... a test of the emergency AOLuser system:-D:-D:-D 2010-09-27T20:49:26 *** bhasker has joined #aichallenge 2010-09-27T20:49:44 @changename penis 2010-09-27T20:49:44 Cyndre_: Error: That operation cannot be done in a channel. 2010-09-27T20:49:56 its against channel nudity 2010-09-27T20:50:01 :P 2010-09-27T20:50:12 * Cyndre_ goes and gets cyndrebot 2010-09-27T20:50:40 *** boiserm has left #aichallenge ("dreaming of planets and ships") 2010-09-27T20:53:01 *** bhasker__ has joined #aichallenge 2010-09-27T20:53:09 *** justin_pdx has quit IRC (Quit: justin_pdx) 2010-09-27T20:53:10 *** bhasker_ has quit IRC (Ping timeout: 272 seconds) 2010-09-27T20:55:30 *** bhasker has quit IRC (Ping timeout: 240 seconds) 2010-09-27T20:57:48 Top 10 players: dmj111(3683), adumlah(3510), davidjliu(3483), sighbot(3465), sequoh(3462), Rexxar(3434), kcm1700(3422), vinx33(3416), McLeopold(3384), Loks1982(3381) 2010-09-27T20:59:58 *** Sap has joined #aichallenge 2010-09-27T21:00:29 I need help with java client :( Can anyone help me? :( 2010-09-27T21:02:11 hello? 2010-09-27T21:02:57 just ask your question, if someone can help you they will respond 2010-09-27T21:03:06 oh, thanx 2010-09-27T21:04:25 *** Sylph has joined #aichallenge 2010-09-27T21:04:37 This game is making me addicted. :/ 2010-09-27T21:05:02 If bot is written on Java it must be single file? can I write a java-bot with several files? 2010-09-27T21:05:43 it can be any number of files but they cannot be in packages 2010-09-27T21:05:45 my bot is great but it has 5 files :( 2010-09-27T21:05:51 *** troutwine has quit IRC (Ping timeout: 265 seconds) 2010-09-27T21:06:10 thanx - I used package :) 2010-09-27T21:06:22 some refactoring is needed :) 2010-09-27T21:06:28 I've been too lazy to bother refactoring... >.> 2010-09-27T21:06:34 yep, then you should be fine 2010-09-27T21:06:34 Im busy rewritting my bot 2010-09-27T21:06:35 good luck :) 2010-09-27T21:07:19 http://pastebin.com/T6fUd0fU thats how I plan to track the game 2010-09-27T21:08:11 anyone think of something I am missing? 2010-09-27T21:09:48 @channel 2010-09-27T21:09:48 Cyndre_: (channel [] []) -- If is given, sets the channel configuration variable for to for . Otherwise, returns the current channel configuration value of . is only necessary if the message isn't sent in the channel itself. 2010-09-27T21:11:32 why i can't see enemy's Fleet DestinationPlanet it my bad or it game rule ? 2010-09-27T21:12:11 you should be able to see any fleets destination 2010-09-27T21:12:20 the game gives you that info 2010-09-27T21:13:19 *** Sap has left #aichallenge 2010-09-27T21:13:55 2 hours spend to find first error into StarterPackage 2010-09-27T21:14:04 now need to find second :( 2010-09-27T21:14:12 :) 2010-09-27T21:15:21 * Naktibalda is trying to find a bug in future orders calculation 2010-09-27T21:15:44 did I miss anything in gameState? 2010-09-27T21:17:54 dunno if it you but in php StarterPackage in Fleet::__construct($owner, $num_ships, $source_planet, $destination_planet, $total_trip_length, $turns_remaining) didn't have this line "$this->destination_planet = $destination_planet;" but must have 2010-09-27T21:19:25 *** bhasker__ has quit IRC (Ping timeout: 264 seconds) 2010-09-27T21:22:05 ups, that's my bug :) 2010-09-27T21:22:15 where did you download it from? 2010-09-27T21:23:33 http://ai-contest.com/forum/viewtopic.php?f=18&t=434 2010-09-27T21:23:51 but anyway DestinationPlanet() return nothing 2010-09-27T21:24:26 download from issue: http://code.google.com/p/ai-contest/issues/detail?id=104 2010-09-27T21:25:07 I fixed some bugs later: + I didn't check what improvements where done by lcsoft 2010-09-27T21:27:25 oh god, Im starting to rewrite my bot from scratch 2010-09-27T21:27:34 * Cyndre_ slaps self 2010-09-27T21:28:50 now i will delete all files which made last 2 days, after that im killing myself and after respawn will start a new bot ;) 2010-09-27T21:32:53 it's pretty addicting, Cyndre_ 2010-09-27T21:33:13 should be doing other things, but i keep reloading to see how my latest iteration of bot is doing 2010-09-27T21:33:33 Sylph: well my bot only plays one turn at a time and makes some generalizations with turns - time to just make it exact 2010-09-27T21:33:37 xah 2010-09-27T21:33:58 http://ai-contest.com/profile.php?user_id=3911 2010-09-27T21:34:16 ah 2010-09-27T21:34:30 http://ai-contest.com/profile.php?user_id=9697 is mine, i just resubmitted 2010-09-27T21:34:43 already fixed a minor bug, but not enough to make me resubmit 2010-09-27T21:35:00 I noticed that the top ranked aren't playing like us, cyndre 2010-09-27T21:35:11 ? 2010-09-27T21:35:26 they build up their forces at key planets 2010-09-27T21:35:34 rather than expand and expand 2010-09-27T21:35:59 a style of top bots is changing constantly 2010-09-27T21:36:03 veery true 2010-09-27T21:36:08 Slyph: it depends on game state 2010-09-27T21:36:08 that's the point of this competition after all 2010-09-27T21:36:14 yeah 2010-09-27T21:36:22 I don't have game state set up yet 2010-09-27T21:36:25 Sylph: once they have more growthRate they send to closest planet to enemy 2010-09-27T21:36:31 Is the tcp server thing down right now? 2010-09-27T21:36:43 yep i noticed that, cyndre 2010-09-27T21:36:48 i stabilize at 300 last time. should get to 250 this time. 2010-09-27T21:36:51 znutar: no 2010-09-27T21:36:54 *** bhasker has joined #aichallenge 2010-09-27T21:37:00 I might bother adding game state and stuff sometimes, 2010-09-27T21:37:04 but my last bot was made to win - not have more ships in 200 turns (rather small time frame) 2010-09-27T21:37:08 But I want to rewrite the entire code. :/ 2010-09-27T21:37:21 but got no time 2010-09-27T21:37:39 I'm somewhat curious how my bot would do against yours 2010-09-27T21:37:39 mm 2010-09-27T21:37:42 but yeah 2010-09-27T21:37:48 Naktibalda: thanks, I guess I'm doing something wrong! 2010-09-27T21:37:54 Cyndre_: don't need to have more growth rate before sending forces to the frontlines 2010-09-27T21:37:55 Sylph: you have 2 months left 2010-09-27T21:37:59 I just started yesterday 2010-09-27T21:38:05 it's a very amusing game 2010-09-27T21:38:09 and time consuming 2010-09-27T21:38:15 znutar: IP was changed a week ago, if you was away for some time :) 2010-09-27T21:38:25 took 6 hours trying to get a visualizer.... ended up switching to java from c# >.> 2010-09-27T21:38:29 coz I wanted my visualizer 2010-09-27T21:38:53 sigh: it also depends on ship count 2010-09-27T21:39:01 naturally 2010-09-27T21:39:02 not really 2010-09-27T21:39:07 Cyndre_: it's just a consequence of the fact that ships are better positioned closer to your enemy than away 2010-09-27T21:39:18 yep 2010-09-27T21:39:21 it's useful for defense 2010-09-27T21:39:28 or attacking 2010-09-27T21:39:34 regardless of who's ahead or behind 2010-09-27T21:39:50 if your forces are at the back you can't attack or defend effectively 2010-09-27T21:39:51 so you just send them to your most central planet? 2010-09-27T21:39:52 noticed it helped a lot on defense for some of the top ranked 2010-09-27T21:39:56 Cyndre_: no 2010-09-27T21:40:05 map central? 2010-09-27T21:40:08 is what I meant 2010-09-27T21:40:34 my bot plays like me lol 2010-09-27T21:40:44 not necessarily, just that your forces should be in the areas where they can attack/defend against the enemy the best 2010-09-27T21:40:52 that is not necessarily one planet 2010-09-27T21:41:03 yeah 2010-09-27T21:41:20 but is definitely not at the back 2010-09-27T21:41:29 you *must* move ships forward, and fast 2010-09-27T21:41:33 like a recording of dmj split his force into 3 planets after he attacked 2010-09-27T21:41:37 and got 4 planets 2010-09-27T21:41:43 otherwise your opponent can easily advance 2010-09-27T21:41:55 yea 2010-09-27T21:42:00 thats probably my main issue 2010-09-27T21:42:12 I don't bother doing that myself yet 2010-09-27T21:42:18 my bot is simple 2010-09-27T21:42:27 what's best to attack or defend, attack or defend them 2010-09-27T21:42:28 don't get me wrong, this is easy enough to say, but tricky to put into practice 2010-09-27T21:42:41 start with the gameState Sylph, you will have to eventually 2010-09-27T21:42:46 I should 2010-09-27T21:42:49 was lazy 2010-09-27T21:42:52 i wanted to put one in 2010-09-27T21:42:56 then i kept adding to it 2010-09-27T21:43:00 sigh: nah, will have to be based on grouping of ships and distance between them 2010-09-27T21:43:11 I have queued stuff I need to do on a total rewrite 2010-09-27T21:43:13 thats what I did 2010-09-27T21:43:14 lmao 2010-09-27T21:43:16 like gamestate 2010-09-27T21:43:33 lol 2010-09-27T21:43:39 so instead of trying to add state later on, Im working on gameState right now 2010-09-27T21:43:39 Cyndre_: sorry, I don't understand 2010-09-27T21:43:43 and then Ill build on top of it 2010-09-27T21:44:03 sigh: based on the distance between my planets and the enemys and the ships at enemys and on my planet 2010-09-27T21:44:25 sigh: my bot loses on 2 rage bot maps for the same reason (doesnt make sure it can defend against ships in range) 2010-09-27T21:44:32 i see, yeah 2010-09-27T21:44:35 but you know, the game state have somewhat limited use to some level. 2010-09-27T21:44:36 sigh: litterly 2 ships short 2010-09-27T21:44:39 :) 2010-09-27T21:44:43 i mean 2010-09-27T21:44:57 if you keep ahead of the enemy from the beginning 2010-09-27T21:45:00 and keep them in check 2010-09-27T21:45:04 you can easily get rank 300s 2010-09-27T21:45:11 I do most of my redistribution based purely on distance 2010-09-27T21:45:14 or probably 200-300s 2010-09-27T21:45:19 I was ranked 150 yesterday 2010-09-27T21:45:25 Im going for double digits 2010-09-27T21:45:28 ah 2010-09-27T21:45:34 yeah i might consider it 2010-09-27T21:45:42 once you've stopped expanding to neutrals, the positions of planets seem to matter much more than growth rate 2010-09-27T21:45:42 been trying to add state and my bots got worse, thus the rewrite 2010-09-27T21:45:46 lol 2010-09-27T21:46:52 it's really aggravating how many matches end any meaningful interaction around turn 50 and then just consist of people building up... 2010-09-27T21:47:50 its a waste to expand glutonously to neutrals 2010-09-27T21:47:50 glutonously, yes 2010-09-27T21:47:50 Cyndre_: agree, it is a balance 2010-09-27T21:47:50 but not always 2010-09-27T21:47:50 i test new iteration of bots on older bots first, and dump changes that make it worse :p 2010-09-27T21:47:50 the balance is between your ships, your growth, and his ships, his growth 2010-09-27T21:47:50 lol davidjliu 2010-09-27T21:47:50 rank 1-100? 2010-09-27T21:47:50 davidjliu: all my local tests against different versions of my bots have started doing this :( 2010-09-27T21:47:50 that don't sound like rank 300-1000 matches 2010-09-27T21:47:50 it makes it really hard to test 2010-09-27T21:47:50 or rank 200-300 2010-09-27T21:47:50 ow sigh 2010-09-27T21:47:50 what rank are you now ? 2010-09-27T21:47:50 davidjliu: I wish it was to the death, I personally dont feel 200 turns is enough to do anything 2010-09-27T21:47:50 @rankings 2010-09-27T21:47:50 @rankings 2010-09-27T21:47:50 eh :p 2010-09-27T21:47:50 tricked me 2010-09-27T21:47:51 davidjiliu: if you waste ships expanding even with double the growth you might now catch up in a short time frame 2010-09-27T21:47:52 sigh: Top 10 players: dmj111(3692), adumlah(3521), davidjliu(3501), sequoh(3489), sighbot(3475), Rexxar(3441), kcm1700(3428), McLeopold(3395), vinx33(3389), Loks1982(3369) 2010-09-27T21:47:53 Sylph: Top 10 players: dmj111(3692), adumlah(3521), davidjliu(3501), sequoh(3489), sighbot(3475), Rexxar(3441), kcm1700(3428), McLeopold(3395), vinx33(3389), Loks1982(3369) 2010-09-27T21:47:57 now = not 2010-09-27T21:48:07 you're one of them ? 2010-09-27T21:48:10 hmmm... #5 apparently 2010-09-27T21:48:13 ah 2010-09-27T21:48:16 nice 2010-09-27T21:48:20 i didn't notice 2010-09-27T21:49:01 got some lucky wins :) 2010-09-27T21:49:01 my bot wasting time fighting low ranked -_- 2010-09-27T21:49:07 taking time to get to its previous rank 2010-09-27T21:49:07 this is very true... my bot just likes the idea of hoarding ships 2010-09-27T21:49:11 which annoys me about uploading 2010-09-27T21:49:16 new bots 2010-09-27T21:49:42 mine doesn't :/ 2010-09-27T21:49:49 http://ai-contest.com/visualizer.php?game_id=4867246 2010-09-27T21:50:09 that's the latest fight >.> 2010-09-27T21:50:10 davidjliu: mine's more stupid, mine likes to hoard ships, and it doesn't even have any idea when it is behind 2010-09-27T21:50:24 so it hoards even when doing so will obviously lose >< 2010-09-27T21:50:27 ah :/ 2010-09-27T21:50:27 go update it 2010-09-27T21:50:28 ? 2010-09-27T21:51:02 ha 2010-09-27T21:51:02 vs: 2010-09-27T21:51:02 http://ai-contest.com/visualizer.php?game_id=4866116 2010-09-27T21:51:02 heh, mine does too... 2010-09-27T21:51:02 I actually don't do aynthing with the current gamestate 2010-09-27T21:51:02 easier said than done :) 2010-09-27T21:51:03 yea 2010-09-27T21:51:03 i want a bot that write itself 2010-09-27T21:51:03 so i can do other things while it does that 2010-09-27T21:51:03 but yeah, that's on the todo list 2010-09-27T21:51:03 I have to refactor my entire code to not be based on the dumb starter package 2010-09-27T21:51:05 Right now its pretty dumb 2010-09-27T21:51:29 ROFLMAO davidjliu 2010-09-27T21:51:32 that's funny 2010-09-27T21:51:55 I se your code send your ships to the closest planet that's closest to enemy. 2010-09-27T21:51:57 from each planet 2010-09-27T21:52:08 vinx33 send theirs straight to the middle 2010-09-27T21:52:17 why are you hoarding lol 2010-09-27T21:52:22 hrm 2010-09-27T21:52:37 yeah... enemy ships scare me a lot, I think 2010-09-27T21:52:38 you got more units 2010-09-27T21:52:40 lol 2010-09-27T21:52:46 my bot is overly defensive 2010-09-27T21:52:52 I'm pretty sure I wait till I have twice as many ships... 2010-09-27T21:52:56 lol 2010-09-27T21:53:26 bacck to 321, wonder how far that little bit of code will push it, hrm. 2010-09-27T21:53:51 ha sigh, concrete example: http://ai-contest.com/visualizer.php?game_id=4824268 2010-09-27T21:53:54 davidjliu?: Im looking at that and planning on using my tot ships - his tot ships as ships for taking neutral planets when Im in the lead 2010-09-27T21:54:30 maybe consider adding some allowance, cyndre. 2010-09-27T21:54:32 ha, the problem with taking neutral planets is always that you might pay for the startup and the enemy just takes it... 2010-09-27T21:54:33 *** bhasker has quit IRC (Read error: Connection reset by peer) 2010-09-27T21:54:38 lol 2010-09-27T21:54:39 yeah 2010-09-27T21:54:46 davidjliu: yeah, I've seen that game... my bot breaks the stalemate 2010-09-27T21:54:57 I call that the suckerpunch 2010-09-27T21:55:08 lol 2010-09-27T21:55:11 i still need to finish watching it 2010-09-27T21:55:25 if my bot hadn't done anything, it would have gotten ahead on growth >< 2010-09-27T21:55:36 lol 2010-09-27T21:55:54 ha, yeah 2010-09-27T21:56:01 lol i see yeah 2010-09-27T21:56:04 david stole that planet 2010-09-27T21:56:07 lol 2010-09-27T21:56:13 in the middle 2010-09-27T21:56:15 that really hurts 2010-09-27T21:56:21 :) 2010-09-27T21:56:49 i'm so curious how bad my bot will do against you guys 2010-09-27T21:56:50 :| 2010-09-27T21:56:59 stupid slow contest site 2010-09-27T21:57:03 I beat amund on tcp server 2010-09-27T21:57:04 ooo 2010-09-27T21:57:14 I'm pretty sure I had a four hour gap between my last two games 2010-09-27T21:57:16 lol 2010-09-27T21:57:21 i dled the tcp server 2010-09-27T21:57:22 i should try it 2010-09-27T21:57:27 http://ai-contest.com/profile.php?user_id=3911 2010-09-27T21:57:29 *** bhasker has joined #aichallenge 2010-09-27T21:57:38 I played 5 games in the last hour..... 2010-09-27T21:57:50 Top 10 players: dmj111(3694), adumlah(3522), davidjliu(3501), sequoh(3490), sighbot(3476), Rexxar(3443), kcm1700(3427), McLeopold(3395), vinx33(3391), Loks1982(3368) 2010-09-27T21:57:59 9 in the last 2 hours myself hrm 2010-09-27T21:58:04 1.5 hours 2010-09-27T21:58:07 grr 2010-09-27T21:59:57 *** blbrown_win3 has joined #aichallenge 2010-09-27T22:00:17 *** bhasker has quit IRC (Read error: Connection reset by peer) 2010-09-27T22:01:07 *** bhasker has joined #aichallenge 2010-09-27T22:02:10 *** bhasker has quit IRC (Read error: Connection reset by peer) 2010-09-27T22:04:34 http://ai-contest.com/profile.php?user_id=3911 sorry david, 3 more games... 2010-09-27T22:05:04 *** dmj111_ is now known as dmj111 2010-09-27T22:05:07 takin peeps out in 181 spot in 56 turns - good bot :) 2010-09-27T22:05:16 http://ai-contest.com/visualizer.php?game_id=4868025 very nice game 2010-09-27T22:05:58 is 200 turns pretty average for a game. It seems kind of low 2010-09-27T22:05:58 Cyndre_: well done 2010-09-27T22:06:13 Cyndre_: also you should try attacking the closer planets first 2010-09-27T22:06:15 Im rarely hitting 200 turns 2010-09-27T22:06:29 sigh: its an algoritm based on distance, ships, and growth rate :) 2010-09-27T22:06:57 Cyndre_: yeah mine too, but I've found that it should favor distance more than I thgouht 2010-09-27T22:06:58 http://ai-contest.com/visualizer.php?game_id=4867873 deemoowoor if your here it looks like you timed out 2010-09-27T22:07:18 *** bhasker has joined #aichallenge 2010-09-27T22:08:05 timing out sucks :/ 2010-09-27T22:08:21 yea, try and let them know - really horrible 2010-09-27T22:08:28 had to fix some infinite loop a couple times, good thing there's a visualizer. 2010-09-27T22:08:30 yeah 2010-09-27T22:08:45 was a good match and Im kinda currious how the last 58 turns would have played out 2010-09-27T22:08:48 *** u1tr0n has quit IRC () 2010-09-27T22:08:54 won my 8 matches 2010-09-27T22:08:58 then lost to the same person i won against 2010-09-27T22:08:59 lol 2010-09-27T22:09:07 on a different map 2010-09-27T22:09:28 lmao, yea maps make a big diff on this one 2010-09-27T22:09:31 I need to fix the defensive routine. 2010-09-27T22:09:37 my defensive routine is very very very dumb 2010-09-27T22:10:00 But that's going to be a little bit of work :| 2010-09-27T22:10:05 mine just locks ships onto a planet... and that's about it :D 2010-09-27T22:10:09 I started writting reinforcePlanet2 and it showed me why I need better state 2010-09-27T22:10:25 my first reinforcePlanet just sent ships till the enemy landed 2010-09-27T22:10:27 very bad 2010-09-27T22:10:30 lol 2010-09-27T22:10:30 ow 2010-09-27T22:10:33 heh 2010-09-27T22:10:43 fun fact... I'm pretty sure I timed out on three of my last four losses. I was probably going to lose at least two anyways, but... it's just annoying 2010-09-27T22:10:44 that would be bad 2010-09-27T22:10:46 so bad it actually performed better with it turned off 2010-09-27T22:10:51 if you fix that, you should stabilize at 300 or better 2010-09-27T22:10:56 at least 2010-09-27T22:11:00 *** bhasker_ has joined #aichallenge 2010-09-27T22:11:04 Im at 233 now Sylph 2010-09-27T22:11:07 ah nice 2010-09-27T22:11:17 and spicytuna beats me down again 2010-09-27T22:11:19 lol 2010-09-27T22:11:39 his bot seems to play more like mine than yours. 2010-09-27T22:11:44 or not sure 2010-09-27T22:11:49 *** mightybyte has quit IRC (Quit: leaving) 2010-09-27T22:12:04 *** bhasker has quit IRC (Ping timeout: 265 seconds) 2010-09-27T22:12:26 crap loss twice in a row hrm 2010-09-27T22:12:36 oh yea i hate getting put next to enemy 2010-09-27T22:12:40 -_- 2010-09-27T22:12:46 I should put in a state 2010-09-27T22:12:50 yeah, those are quite tricky 2010-09-27T22:12:59 if it start next to each other, just wait and attack enemy base >.> 2010-09-27T22:13:03 or something 2010-09-27T22:13:29 if are are starting to come across that, then try getting your bot to beat ragebot on all 100 starter maps 2010-09-27T22:13:42 98 is my best 2010-09-27T22:13:43 that will help you in close combat 2010-09-27T22:13:52 work on the last two :) 2010-09-27T22:14:00 bugs 2010-09-27T22:14:14 i win everything against those bots. :/ 2010-09-27T22:14:44 not many people use ragebots style - only encounter that 600+ 2010-09-27T22:15:13 Cyndre_: sure, but what you learn from countering ragebot is important 2010-09-27T22:15:33 don't optimise your bot against ragebot, that's not what I'm saying 2010-09-27T22:15:35 :P 2010-09-27T22:15:37 sigh: my biggest issue is timing from what I have seen 2010-09-27T22:15:51 timing of what? 2010-09-27T22:16:11 when ships are going to land - I generalize 2010-09-27T22:16:14 all those basic bots are so easy to beat though sigh :/ 2010-09-27T22:16:23 how many turns Sylph? 2010-09-27T22:16:29 which stage 2010-09-27T22:16:36 all of them 2010-09-27T22:16:39 has been as f ast as 26 turns in a couple 2010-09-27T22:16:51 i can't say all of them xd 2010-09-27T22:16:54 ask which bot and which map 2010-09-27T22:16:55 :p 2010-09-27T22:16:55 I average 55 and lose twice to ragebot on map 22 and 30 2010-09-27T22:17:02 that's nice 2010-09-27T22:17:20 greghaynes wrote a nice script for testing against all of them 2010-09-27T22:17:26 rage bot dies within 34 turn on map 22 for me 2010-09-27T22:18:01 interesting on map 88. seems that's going to end up a draw 2010-09-27T22:18:10 nm, i win barely 2010-09-27T22:18:14 -_- 2010-09-27T22:18:22 42 turns I die on 22 2010-09-27T22:18:22 Thanks for telling me a map that my bot is weak on 2010-09-27T22:18:34 i won on 165 turns, that's so slow :/ 2010-09-27T22:18:58 map 30 I lose on turn 200 2010-09-27T22:19:06 oh wait, that was map 30. on map 88 i lost 2010-09-27T22:19:36 *** bhasker has joined #aichallenge 2010-09-27T22:19:39 30 turn win on map 88 2010-09-27T22:20:10 i lost on 88, that bs hrm 2010-09-27T22:20:15 I hate waiting for match's 2010-09-27T22:20:24 same 2010-09-27T22:20:56 Sylph: have you tried all maps? 2010-09-27T22:21:05 no, but that gets tiring 2010-09-27T22:21:09 and takes a while 2010-09-27T22:21:10 :) 2010-09-27T22:21:10 with Java, does the system support java packages when I upload? 2010-09-27T22:21:37 *** bhasker_ has quit IRC (Ping timeout: 265 seconds) 2010-09-27T22:21:46 Sylph: check github for greghaynes and he has a python script for testing against all maps 2010-09-27T22:21:47 Sylph: I'm using this script: http://gist.github.com/574904 2010-09-27T22:21:52 ah 2010-09-27T22:21:55 ty Naktibalda 2010-09-27T22:21:57 thanks 2010-09-27T22:23:05 i still need to try out the tcp too 2010-09-27T22:23:06 and stuff 2010-09-27T22:23:20 i just started doing this yesterday >.> 2010-09-27T22:24:53 Nah, I dont have the python script, I have the GUI for the tcp server 2010-09-27T22:25:03 Rather, the real time game visualizer 2010-09-27T22:25:16 I just use bash + find to test against all maps/bots 2010-09-27T22:25:23 * Sylph nods 2010-09-27T22:25:56 oh, dimitri has the script 2010-09-27T22:25:58 sorry 2010-09-27T22:26:10 btw - awsome job on tcp server greghaynes 2010-09-27T22:26:12 er client 2010-09-27T22:26:14 ty 2010-09-27T22:27:09 mm, you got a gui tcp client one ? 2010-09-27T22:27:11 where is it xD 2010-09-27T22:27:15 ty ty 2010-09-27T22:27:21 Sylph: Yep, real time shows games 2010-09-27T22:27:26 serious? 2010-09-27T22:27:34 http://github.com/greghaynes/QTcpPlanetWars 2010-09-27T22:27:36 it get tiring to shift+right click to open command line 2010-09-27T22:27:44 *** jimmerton has quit IRC (Quit: jimmerton) 2010-09-27T22:28:12 ah, no binary version? 2010-09-27T22:28:13 * Sylph download 2010-09-27T22:28:41 *** dmj111 has quit IRC (Quit: ERC Version 5.3 (IRC client for Emacs)) 2010-09-27T22:28:53 If someone wants to make one, I havent learned to package before 2010-09-27T22:28:59 People always do it for me ;) 2010-09-27T22:29:15 *** dmj111 has joined #aichallenge 2010-09-27T22:29:15 :) 2010-09-27T22:30:10 that python thing takes a while 2010-09-27T22:30:36 it's 100 games per bot, try to do it faster manually :) 2010-09-27T22:30:45 :p 2010-09-27T22:30:50 not sure if its edited, but if you open test.py and ctrl-f 1000 1000 replace with 200 200 it will be like the game 2010-09-27T22:30:54 yeah 2010-09-27T22:31:10 I'm not sure what to do if someone is next to me, hrm. 2010-09-27T22:31:14 just wait them out and attack them ? :/ 2010-09-27T22:32:10 Sylph: seems how you brought it up - MyPlanetShips + (MyPlanetGrowth * distance(MyPlanet, Enemy)) - EnemyShips = avalships 2010-09-27T22:32:21 *** jimmerton has joined #aichallenge 2010-09-27T22:32:38 True I can do that hrm 2010-09-27T22:33:18 and then it gives you some ships to send out the first turn, and if the enemy dedicates them all you can either take main planet or dedicate more to expansion 2010-09-27T22:33:25 * Sylph nods 2010-09-27T22:33:25 *** jimmerton has quit IRC (Client Quit) 2010-09-27T22:34:03 good idea 2010-09-27T22:34:13 I think I'm going to move availableship to an array 2010-09-27T22:34:18 from what i have it atm 2010-09-27T22:34:27 you mentioned it which gave me the idea on how to fix my last two maps ragebot beats me up 2010-09-27T22:34:33 yeah 2010-09-27T22:34:36 Cyndre_: 200 200 is a 200 move game with 200ms timeout 2010-09-27T22:34:51 sigh: lmao 2010-09-27T22:35:01 my bot doesnt timeout on my home machine 2010-09-27T22:35:03 http://pastebin.com/T6fUd0fU 2010-09-27T22:35:04 guess not, no java packages. http://ai-contest.com/forum/viewtopic.php?f=18&t=731&p=4582&hilit=packages#p4582 2010-09-27T22:37:05 and from rank 224 I descend 2010-09-27T22:37:42 to what position? 2010-09-27T22:37:57 my bot dropped to 156, before upload 2010-09-27T22:38:03 227 2010-09-27T22:38:24 I hope that the new version will climb to top50 :) 2010-09-27T22:39:41 I think i'll just add that routine to my current defensive routine, cyndre. 2010-09-27T22:39:47 hrm 2010-09-27T22:39:53 and loop through enemy planets >.> 2010-09-27T22:39:56 this is so wasting resources 2010-09-27T22:40:19 Sylph: I have a ship dispatcher that makes sure I dont do bad things 2010-09-27T22:40:22 lol 2010-09-27T22:40:35 I've been lazy 2010-09-27T22:40:36 Sylph: I have loops like that all over the place :D 2010-09-27T22:40:38 just added up to that one function 2010-09-27T22:40:44 and forget stuff 2010-09-27T22:41:03 that would make my code nicer looking and modular 2010-09-27T22:41:06 okay then sigh :p 2010-09-27T22:41:24 I copied the data on enemy planets, etc 2010-09-27T22:41:42 so I don't have to keep requesting:/ 2010-09-27T22:42:23 I also found it handy to compute stuff like planets sorted by distance once and store it, because i keep using it 2010-09-27T22:43:02 *** j3camero has joined #aichallenge 2010-09-27T22:43:05 yeah 2010-09-27T22:43:10 I did that for fleet, etc 2010-09-27T22:43:12 yo yo 2010-09-27T22:43:14 i should keep track of fleet arrival time 2010-09-27T22:43:17 but like cba atm 2010-09-27T22:43:34 I figured it's about time I checked back into the IRC channel. 2010-09-27T22:43:44 long time no see 2010-09-27T22:43:46 hi! 2010-09-27T22:44:13 uhoh, bix0r's coming back... 2010-09-27T22:44:14 So how are people finding it nowadays? Any major peeves? 2010-09-27T22:44:23 yea 2010-09-27T22:44:28 my bot keeps losing to the top 4 or so 2010-09-27T22:44:29 my bot sucks 2010-09-27T22:44:31 lmao 2010-09-27T22:44:31 lol 2010-09-27T22:44:32 same 2010-09-27T22:44:33 can you fix that? 2010-09-27T22:44:40 lol 2010-09-27T22:44:46 my bot keeps timing out. aww 2010-09-27T22:44:49 its been like 7 minutes since my last game 2010-09-27T22:44:51 sigh: yeah no problem just gimme a sec. 2010-09-27T22:44:52 lmao 2010-09-27T22:45:21 awesome 2010-09-27T22:45:23 screw that, I keep losing to top 40 or so, I think my fix gets priority over sigh 2010-09-27T22:45:30 INSERT INTO games (winner,loser) VALUES ("sigh",rand()); X 100 2010-09-27T22:45:33 I've had one game in the last 5 hours, almost 2010-09-27T22:45:36 That ought to do it. 2010-09-27T22:45:40 hhaa 2010-09-27T22:45:44 lol 2010-09-27T22:46:09 could you, like, actually fix that? 2010-09-27T22:46:42 davidjliu: yep, it's a big issue that I am spending a lot of time worrying about. It seems like there is a lot of timing out going on lately. 2010-09-27T22:46:55 is that just because of server load? 2010-09-27T22:47:09 Are the timeouts spread evenly across all the bots, or is it just some languages more than others? 2010-09-27T22:47:21 sigh: the cause is unknown. Probably just my shitty code. 2010-09-27T22:47:28 ah :( 2010-09-27T22:47:45 aw... 2010-09-27T22:48:08 http://ai-contest.com/visualizer.php?game_id=4870130 wtf 2010-09-27T22:48:21 lol 2010-09-27T22:48:26 lol 2010-09-27T22:48:30 haha I could probably fix it in a day, but I am starting a new full-time job. If it wasn't for the other volunteers, there would be very little contest to speak of, haha. 2010-09-27T22:49:10 despite the timeouts, are you finding you're able to get a more or less fair ranking? Or are ALL your games timing out? 2010-09-27T22:49:10 funny switching, Cyndre_ 2010-09-27T22:49:19 none of my games are timing out xD 2010-09-27T22:49:36 Sylph: which language? 2010-09-27T22:49:39 hmm 2010-09-27T22:49:40 wait 2010-09-27T22:49:46 i use java atm 2010-09-27T22:49:47 I've had, like, three of my games time out, I think. 2010-09-27T22:49:48 let me check last 12 matches 2010-09-27T22:49:57 davidjliu: out of how many? 2010-09-27T22:50:31 j3camero: zomg, you're on irc 2010-09-27T22:50:35 actually, it was 4, but out of a lot 2010-09-27T22:50:40 amstan: oh snap! 2010-09-27T22:50:44 one of them was a long time ago and I think that one was probably legit 2010-09-27T22:50:51 but I just had three in quick succession today 2010-09-27T22:51:00 davidjliu: oh that is good news. I officially care much less now (in favor of other issues, of course) 2010-09-27T22:51:09 okay, http://ai-contest.com/visualizer.php?game_id=4867246 is timed out ? 2010-09-27T22:51:29 amstan: sorry I am not spending as much time on the contest as I should be. I got drafted to be the CTO of a political campaign. 2010-09-27T22:51:29 only 87 turns, enemy last planet hasn't turned color 2010-09-27T22:51:40 my bot always strategically times out when it has no planets left 2010-09-27T22:51:44 how do you know if things time out 2010-09-27T22:51:50 lol david 2010-09-27T22:51:55 i mean, other than it ending half way 2010-09-27T22:52:22 like some games end with enemy planets all gone 2010-09-27T22:52:25 davidjliu: your bot errors when there are no planets? mine was doing that before I fixed 2010-09-27T22:52:27 some seems to have the last planet still 2010-09-27T22:52:32 davidjliu: if planets.length() >= 1 :) 2010-09-27T22:52:50 yeah, I don't really care, cuhz I'm probably going to lose anyways 2010-09-27T22:53:04 and I have no idea why I spelled that cuhz... 2010-09-27T22:53:05 davidjliu: username? 2010-09-27T22:53:18 I use for (Planet p : myPlanets) { :/ 2010-09-27T22:53:27 j3camero: he's just davidjliu, he's #3 :P 2010-09-27T22:53:31 so if it don't exist well, it don't go anyway :| 2010-09-27T22:53:31 mm 2010-09-27T22:53:34 makes sense. 2010-09-27T22:53:36 or does java not work like that 2010-09-27T22:53:47 nm, it s same thing 2010-09-27T22:53:56 so i don't get why you'd get time out >.< 2010-09-27T22:54:14 davidjliu: are you finding it's competitive at the top of the leaderboard? Like, do you really believe that your bot is solildly in third place? Or is there a bit of random noise there? 2010-09-27T22:54:24 um... I think it's pretty random a lot 2010-09-27T22:54:35 as in, I was like six earlier I think 2010-09-27T22:54:40 and then I sat there for a few hours 2010-09-27T22:54:42 I think dmj111 is a pretty clear first 2010-09-27T22:54:47 (without playing any games) 2010-09-27T22:54:51 yeah somehow, 2010-09-27T22:54:53 and I think I just became third 2010-09-27T22:54:55 I need to rewrite my entire code anyway hrm 2010-09-27T22:55:03 maybe i'll forget fixing this bot 2010-09-27T22:55:05 and just rewrite :| 2010-09-27T22:55:27 like I've been third for a while, but it's just cause I haven't played anyone in a while 2010-09-27T22:55:31 davidjliu: my biggest fear is that the top 10 have basically found "the best strategy" and that there are very few gains left to have. Can someone from the top address this concern? Are there still lots of improvements to be had? 2010-09-27T22:55:37 each game between teams at the top change a alot 2010-09-27T22:55:38 j3camero: no way 2010-09-27T22:55:43 I think there are still lots of improvements, yeah... 2010-09-27T22:55:48 yeah 2010-09-27T22:55:54 i saw dmj complaining about a lot of things they are going to fix 2010-09-27T22:55:56 in the forum :p 2010-09-27T22:55:59 yeah 2010-09-27T22:56:04 about their bot 2010-09-27T22:56:10 no one still has a clear idea of what is best 2010-09-27T22:56:19 That's great news! 2010-09-27T22:56:33 I'm glad Phil Hassey found me and pitched Galcon as the game for this contest. 2010-09-27T22:56:37 :) 2010-09-27T22:56:48 yeah, it's a good game 2010-09-27T22:56:51 nice choice :) 2010-09-27T22:56:59 yeah 2010-09-27T22:57:04 I really had no idea if it was going to be "deep" enough. You know what I mean? 2010-09-27T22:57:04 I'm really enjoying it :D 2010-09-27T22:57:11 yeah 2010-09-27T22:57:18 the best games are often the simplest or vice versa. :p 2010-09-27T22:57:25 would have been more confusing if you go do eufloria :p 2010-09-27T22:57:30 i think there is still a way to go to find the _best_ strategy... but I do wonder how long it will take someone to find it. 2010-09-27T22:57:35 honestly, the fact that we had no idea whether it would be deep was compelling to me 2010-09-27T22:57:38 j3camero: i'm pretty unsatisfied with the state of stuff, there's lots of things to be done, so many things fail to work 2010-09-27T22:57:40 so far, I haven't been able to apply many "traditional" AI techniques 2010-09-27T22:57:48 this honestly requires thinking about strategy a lot 2010-09-27T22:57:49 Top 10 players: dmj111(3678), adumlah(3510), davidjliu(3489), Rexxar(3479), sequoh(3477), sighbot(3451), bix0r4ever(3429), kcm1700(3410), McLeopold(3378), vinx33(3378) 2010-09-27T22:57:54 people are still complaining about the cpu time vs wall time, we did not get an answer for that 2010-09-27T22:58:05 I made the shallowness mistake a while back when we did a Rock Paper Scissors contest. It got to the point where all the top bots were the same basically. There is only so good you can get at RPS. 2010-09-27T22:58:19 oh yeah, and is it still a rule that timing out results in a suspension from the official competition? 2010-09-27T22:58:39 RPS is very well researched, too 2010-09-27T22:58:58 davidjliu: no, timeouts just make you lose the game. 2010-09-27T22:59:11 like, in the official contest too? cause apparently it was different for tron... 2010-09-27T22:59:11 j3camero: that's good 2010-09-27T22:59:34 * Naktibalda just dropped down by 107 positions without playing any game 2010-09-27T22:59:36 amstan: yeah, I hear that. Sorry again for being AWOL. I am lucky to even get this half-hour to chat in IRC. I am swamped. 2010-09-27T23:00:06 j3camero: there's still lots of things to do, it's not at the point where it can run itself 2010-09-27T23:00:15 amstan: I know. 2010-09-27T23:00:35 most things require your decisions 2010-09-27T23:00:41 but like.. look at the forums 2010-09-27T23:00:44 okay 2010-09-27T23:00:55 it's full of people just complaining about stuff 2010-09-27T23:01:11 about how they can't change their profile, and how they timeout inexplicably, or how it's all slow 2010-09-27T23:01:28 I was looking about changing profile and yeah noticed the rare matches 2010-09-27T23:01:42 not to mention password reset requests 2010-09-27T23:01:47 emails for the forums still don't work 2010-09-27T23:02:09 making every little guy who forgets his password to try to contact anyone and post in as many places as possible 2010-09-27T23:03:20 > Sylph : you saw me complaining in the forums? maybe someone else? 2010-09-27T23:03:42 *** Naktibalda has quit IRC (Remote host closed the connection) 2010-09-27T23:04:36 amstan: yep. As usual, the contest pretty much sucks, hehe. 2010-09-27T23:04:39 not me 2010-09-27T23:04:43 i haven't posted in the foru 2010-09-27T23:04:44 m 2010-09-27T23:04:46 actually, it was this guy that pitched galcon first: http://eu.reddit.com/r/programming/comments/bxfnk/what_game_should_we_pick_for_the_next_google_ai/c0p1w68 2010-09-27T23:04:49 i barely started doing this yesterday 2010-09-27T23:05:28 I got a large list of code to do to push my code to top 25. :/ 2010-09-27T23:05:32 the contest is wonderful 2010-09-27T23:05:34 unfortunately got a lot of things to do 2010-09-27T23:05:35 amstan: the political campaign thing should wind down in a few days from now. Then I will be able to renew a push to make the contest not suck. We are closer than ever to having a more-or-less flawless contest. It's definitely achievable. 2010-09-27T23:05:38 :| 2010-09-27T23:06:09 Sylph: my bad... i mixed your comment about how I want to fix my code with the discussion about people having problems in the contest. the "complain" word is all I saw. Sorry. 2010-09-27T23:06:11 yeah i have been distracted by many things. i'm so glad the deadline this time around is so much farther away 2010-09-27T23:06:15 ah 2010-09-27T23:06:26 lol 2010-09-27T23:06:34 I saw you mentioning how you wanted to improve your bot on the next iteration or something 2010-09-27T23:06:46 though it might be a few iterations ago 2010-09-27T23:07:05 that was today... i am avoiding trying to figure out what to do next right now :) 2010-09-27T23:07:08 ah 2010-09-27T23:07:12 yeah 2010-09-27T23:07:16 i have assignments to do 2010-09-27T23:07:20 I have to rewrite my entire code -_- 2010-09-27T23:07:32 Need to stuff an actual brain into it 2010-09-27T23:07:38 But need to do other things first 2010-09-27T23:07:57 dmj111: hey, it's nice to see you around! How are you liking (or not liking) the contest? 2010-09-27T23:08:18 hey j3camero .... Of course I won't complain right now :) 2010-09-27T23:08:18 dmj111: you're pretty solidly in first place. Do you think you'll get taken down, or do you have enough new ideas that you can stay on top? 2010-09-27T23:08:26 definitely get taken down. 2010-09-27T23:09:00 dmj111: do you mind telling me a bit about where you're from? University student, full-time software engineer, ...? 2010-09-27T23:09:27 the odds are low that I will win, but I hope to stay in the mix. I have a few ideas that might make a big difference.. but I think over the course of two months other people will top it. 2010-09-27T23:09:55 np, full time programmer. BS and MS in math, lots of interest in computer science. 2010-09-27T23:10:14 Awesome. Which field do you work in? 2010-09-27T23:10:51 I worked for 10 years in defense contractors, and the last two in a commercial company developing business applications. 2010-09-27T23:11:07 * jmcarthur got tired of defense contracting after only two years 2010-09-27T23:11:10 Sounds interesting, haha! 2010-09-27T23:11:19 I guess we dont have much of that in Canada. 2010-09-27T23:11:59 well, I guess the first five were fun because it was at a nice research type place. The next five were fun cause I had a lot of responsibility, but got tired of how things were managed. 2010-09-27T23:12:23 i enjoyed the researchiness of the job i had, but i got tired of the material anyway 2010-09-27T23:13:01 re the contest: i am having a lot of fun. I am amazed at the number of people you have ,and the fact that you are keeping everything moving. hats off to all of the volunteers running this. 2010-09-27T23:13:08 I work in Finance at the moment. I am quite enjoying it. It's very researchy. My job is to sit at my desk all day and play with huge datasets. 2010-09-27T23:13:38 sounds enjoyable 2010-09-27T23:13:46 cool :) 2010-09-27T23:13:53 sounds fun. do you right custom software for that, or use anything in particular.... or can't say :) 2010-09-27T23:14:06 dmj111: are you protecting your strategy, or do you mind sharing the general approach that you're taking? 2010-09-27T23:15:09 dmj111: yeah, I write tons of Python scripts usually. I use very little pre-made software, because most of it is not up to the task, or is very expensive. It boils my blood when a bank pays 100,000 per year for a piece of software that is basically just grep. 2010-09-27T23:15:18 well, I was thinking about writing stuff up eventually (sooner than later to help beginners out)... and unfortunately, it is complex enough I cannot predict what it will do. 2010-09-27T23:15:24 lol j3, thaat's bad. 2010-09-27T23:16:09 yep... I have pushed python pretty hard to my coworkers for years. "you can do that in python" is kind of my catch phrase. 2010-09-27T23:16:23 dmj111: I would seriously encourage you to produce some content, even if it's just a few short blog posts. I think that would be amazing. 2010-09-27T23:16:30 j3camero: so what does one need to do to sell grep to a finance corporation for 100K/year? ;) 2010-09-27T23:16:41 jmcarthur: excellent question.. 2010-09-27T23:16:50 * dmj111 would like to do that... 2010-09-27T23:16:53 jmcarthur: it's obscenely easy, hehe. 2010-09-27T23:16:59 dmj111: i asked first! 2010-09-27T23:17:11 * dmj111 waits his turn... 2010-09-27T23:17:28 lol 2010-09-27T23:17:32 i'll take the next turn then 2010-09-27T23:17:47 it's usually done the standard way. You sell it to managers somehow without consulting any technical people seriously. 2010-09-27T23:17:48 * sigh patents the idea and waits... 2010-09-27T23:17:51 with that kind of 'job', I'd be able to actually work on this fun competition code 2010-09-27T23:18:35 okay, so we just need a complicated gui for grep that will make managers go "wow" 2010-09-27T23:18:52 jmcarthur: don't remember to be full buzzword compliant 2010-09-27T23:19:00 *don't forget 2010-09-27T23:19:01 goes without saying 2010-09-27T23:19:07 grep internet 2010-09-27T23:19:34 is it web-scale? 2010-09-27T23:19:44 lol 2010-09-27T23:19:46 it's cloud-scale 2010-09-27T23:19:51 my company does a pretty good grep O:-) 2010-09-27T23:20:05 dmj111: producing content that is accessible to self-taught hobbyists is a noble cause. Part of the reason I like running these contests is because it makes learning CS fun. At least hundreds of people could be said to be gaining skills through this contest. They bring those skills to their schools and workplaces, and make the world a better place. So I would highly encourage you to create some tuts or other spill-the-beans content. 2010-09-27T23:20:23 here is my strategy in a nutshell... attack neutrals when it is completely safe, look into the future and see if i can take planets away from the enemy. The competing parts of figuring out safety and figuring what I can take is the hard part, and I am not sure how the two always fit together. 2010-09-27T23:20:23 *** antimatroid has quit IRC (Read error: Connection reset by peer) 2010-09-27T23:20:29 I want a c# visualizer, j3camera :/ 2010-09-27T23:20:35 none of that is probably surprising... 2010-09-27T23:20:37 Tempting to write one :| 2010-09-27T23:20:56 but would be nice if someone make it and I can use it :/ 2010-09-27T23:21:03 I had to move to java, rather than have fun with C# -_- 2010-09-27T23:21:26 *** antimatroid has joined #aichallenge 2010-09-27T23:21:29 j3camero: arm twist.... i should start working on it... cause when my bot starts clinking down the ladder, i will be motivated to work on it again. 2010-09-27T23:21:30 Sylph: look around the forums I thought someone had done that, but maybe I'm misremembering 2010-09-27T23:21:34 how is the visualizer influencing your choice of language, Sylph? 2010-09-27T23:21:42 heavily, jmcarthur 2010-09-27T23:21:48 i mean the languages are basically the same anyway... :/ 2010-09-27T23:21:50 i don't understand how 2010-09-27T23:22:04 It boils down to strategy and being able to see how your strategy impact the bot performance.... 2010-09-27T23:22:08 the language the visualizer is written in seems irrelevant to me 2010-09-27T23:22:12 hrm 2010-09-27T23:22:17 I don't know how they work in the interpreting 2010-09-27T23:22:20 * Janzert agrees 2010-09-27T23:22:22 jmcarthur: why would the visualiser affect your language? 2010-09-27T23:22:24 maybe i need to read more on that 2010-09-27T23:22:29 dmj111:haha yeah that's where the most motivated contestants come from. The "fallen winners" who want to get back on top! 2010-09-27T23:22:38 I didn't know if the java visualizer is written to interpret C# bot or not 2010-09-27T23:22:44 sigh: that's what i'm asking 2010-09-27T23:22:44 the visualizer needs nothing but the playback output from the engine 2010-09-27T23:22:50 ah 2010-09-27T23:22:51 Sylph: they communicate via a simple text protocol 2010-09-27T23:22:53 the engine needs nothing but stdout from the bots 2010-09-27T23:22:53 j3camero: it works the other way too, you can lose motivaton that way 2010-09-27T23:22:56 ah ok 2010-09-27T23:22:58 that's easy then 2010-09-27T23:23:02 jmcarthur: ah, sorry misunderstood you 2010-09-27T23:23:04 Thanks 2010-09-27T23:23:09 I might go back to c# 2010-09-27T23:23:10 *** Olathe has quit IRC (Ping timeout: 245 seconds) 2010-09-27T23:23:16 yes, they may be written in java, but the bots can be written in anything 2010-09-27T23:23:32 yeah, if they just need stdouts and stuff 2010-09-27T23:23:37 I didn't know how they were written 2010-09-27T23:23:39 There was a guy from Poland in the Tron contest who was on top for weeks, then he lost the top spot. His submission rate noticeably increased. 2010-09-27T23:23:47 :) 2010-09-27T23:23:59 I guess reading the bot file and stdin/stdout it's pretty obvious now... 2010-09-27T23:24:03 >.> 2010-09-27T23:24:14 thankts 2010-09-27T23:24:22 tron is kind of a blur for me now...the last few weeks i messed everything up. 2010-09-27T23:24:24 np 2010-09-27T23:24:24 I barely started this yesterday, so I'm still not oriented 2010-09-27T23:24:31 Anyways, I have to stop neglecting my girlfriend at this point. I am glad people are enjoyin the contest! Talk soon. 2010-09-27T23:24:43 later. 2010-09-27T23:24:44 bye! 2010-09-27T23:24:49 j3camero: if you still need people I'm still willing to help sort issues and such. 2010-09-27T23:24:59 oops, cya 2010-09-27T23:25:09 j3camero: thanks for wrestling some secrets out of dmj :P 2010-09-27T23:25:12 Janzert: yes. Send me your gmail address in a PM. Imma add you. 2010-09-27T23:25:27 where :D 2010-09-27T23:25:38 * Sylph scrolls up in case the wrestling worked 2010-09-27T23:25:57 Ah I see 2010-09-27T23:26:13 yeah, future prediction was something I was considering to do on my rewrite. :/ 2010-09-27T23:26:31 * Sylph keep that noted 2010-09-27T23:27:32 i noticed you only ever attack with a full fleet that's more than enough to take enemy planet, dmj 2010-09-27T23:27:44 considering nearby defenses and stuff 2010-09-27T23:29:11 Sylph: he might just be doing that because it is easier, only one choice to make: do I have enough ships, yes or no? 2010-09-27T23:29:15 yeah 2010-09-27T23:29:45 *** bhasker has quit IRC (Read error: Connection timed out) 2010-09-27T23:29:48 it's better from the perspective that it keeps enemy guessing til last second 2010-09-27T23:29:57 that too :) 2010-09-27T23:30:16 my bot would really misguess because of that. 2010-09-27T23:30:22 if you send a steady stream, you give your opponent time to reorganise 2010-09-27T23:30:34 yeah, and its obvious to them what is going to get attacked, etc 2010-09-27T23:30:50 easy plucking planets from enemy hand 2010-09-27T23:30:54 right after they take it 2010-09-27T23:31:06 yeah, it's the time lag, how quickly you act matters *a lot* 2010-09-27T23:32:14 *** bhasker has joined #aichallenge 2010-09-27T23:34:12 *** keynes has joined #aichallenge 2010-09-27T23:34:18 Yes.... so, my notes above were just kind of the code organization, not so much the ideas behind it. 2010-09-27T23:34:50 One principle I was following was delaying until the last second. but... there are also times when that is not the best plan. 2010-09-27T23:35:18 *** znutar has quit IRC (Quit: alice.) 2010-09-27T23:35:19 dmj111: I do the same, and I also find that it is not *always* the best idea 2010-09-27T23:35:33 sometimes it is better to deprive the opponent of ships, fast 2010-09-27T23:35:43 but I don't know how to tell the difference :( 2010-09-27T23:36:01 probably depends on the opponent 2010-09-27T23:36:26 I mean, I personally think that a good strategy is to, when ahead and close enough to the opponent, send a constant stream of ships at all times 2010-09-27T23:36:30 I don't actually do this though... 2010-09-27T23:36:32 yes. one of my development deadends included code to wear the enemy down if I have the ships/growth rate. didn't work the best though. 2010-09-27T23:36:46 davidjliu: i wonder about that too... 2010-09-27T23:37:19 yeah, the problem I have is that introduces a lot more descisions... much more complex 2010-09-27T23:37:28 anyway, Obeleh's writeup in the forum covers most of what I do . ... 2010-09-27T23:37:33 http://ai-contest.com/forum/viewtopic.php?f=17&t=593&start=0 2010-09-27T23:37:46 dmj111: yeah, I often link people to that 2010-09-27T23:37:55 when they ask for basic strategy advice 2010-09-27T23:38:06 dmj111: except you probably have the best implementation too, heh 2010-09-27T23:38:10 sigh: I keep killing myself with complexity... 2010-09-27T23:38:14 everytime I've tried sending a constant stream I end up starving my planets of ships 2010-09-27T23:38:22 dmj111: :) 2010-09-27T23:38:46 davidjliu: thanks. a few more matches with top 20 people, though, and I am sure I will be knocked down several pegs. just waiting for it to happen ;) 2010-09-27T23:38:50 sending ships in one go is much easier to decide 2010-09-27T23:38:51 urg, and I've still played two games in the last five and a half hours... 2010-09-27T23:38:58 ha, you're up 150 points on #2... 2010-09-27T23:39:17 *** keyne has joined #aichallenge 2010-09-27T23:39:19 earlier he was over 200 :P 2010-09-27T23:39:23 thanks for the url, dmj 2010-09-27T23:39:37 that'll be helpful to help keep my rewrite in order 2010-09-27T23:40:12 that may be the luck of the draw. I would like to see the matchups be even closer in ranking, but several of my losses were out of the top 100, or even top 300 I think. 2010-09-27T23:40:24 Sylph: no problem. 2010-09-27T23:40:42 what do you do if you're on the losing end anyway? My idea is just to never be on the losing end... :/ 2010-09-27T23:40:43 dmj111: really? were thy legit losses, or just timeout? 2010-09-27T23:40:49 defeinitely agree with the closer in ranking matchups, at least for the top 20-ish 2010-09-27T23:40:50 if i'm ever in the losing end, just optimize better >.> 2010-09-27T23:40:51 Sylph: ha! 2010-09-27T23:41:01 *** keynes has quit IRC (Ping timeout: 264 seconds) 2010-09-27T23:41:03 *** mondo has joined #aichallenge 2010-09-27T23:41:06 sigh: yep. legit losses. 2010-09-27T23:41:14 let me look... 2010-09-27T23:42:03 huh, they have rolled out of my history. now I understand my rating. 2010-09-27T23:42:08 :) 2010-09-27T23:42:12 dmj111: I mean, I remember looking, and I feel that you generally expand at somewhat close to the rate the opponent expands 2010-09-27T23:42:24 (very generally) 2010-09-27T23:42:28 this one is ranked 780 now, and I cannot remember what he was ranked when we play. 2010-09-27T23:42:35 s/play/played/ 2010-09-27T23:42:37 http://ai-contest.com/profile.php?user_id=7566 2010-09-27T23:43:00 seems simple strategy 2010-09-27T23:43:11 davidjliu: one weakness I have is being overly cautious with neutrals. 2010-09-27T23:43:19 dmj111: the 780 is meaningless, he hasn't lost yet :P 2010-09-27T23:43:35 yeah 2010-09-27T23:43:35 heh, except the thing you do to other people a lot is steal their neutrals... 2010-09-27T23:43:49 was probabl why he said don't know what rank he was before : 2010-09-27T23:43:51 mm 2010-09-27T23:43:57 lol david 2010-09-27T23:44:01 *** keyne has quit IRC (Ping timeout: 264 seconds) 2010-09-27T23:44:20 sigh: yeah.. I pushed it those losses out of my memory :) I am not sure anymore... 2010-09-27T23:44:30 :) 2010-09-27T23:44:43 losses are the best to learn from, should keep them in your memory :P 2010-09-27T23:44:49 davidjliu: yeah.... i guess i am overly cautious because of that. 2010-09-27T23:45:05 I remember my losses to you and davidjliu well :D 2010-09-27T23:45:13 heh 2010-09-27T23:45:19 well, they were today, but still... 2010-09-27T23:45:51 sigh: true. I have two or three known problems that I want to fix. I _think_ they are the causes of my losses... also, I do have them saved off.. but it does not include rankings. 2010-09-27T23:46:08 yeah, fair enough 2010-09-27T23:47:28 Sylph: right now, my code is not aware of whether it is ahead or behind. 2010-09-27T23:47:34 * Sylph nods 2010-09-27T23:47:48 dmj111: I glad I'm not the only one :) 2010-09-27T23:47:49 I'm trying to think of what language to use for my rewrite, mm 2010-09-27T23:48:01 C# would be nice, though shame no C# 4.0 2010-09-27T23:48:10 dmj111: mine isn't aware either... 2010-09-27T23:48:19 i thought about F#, but python is so easy... 2010-09-27T23:48:23 lol 2010-09-27T23:48:34 ruby looks pretty too 2010-09-27T23:48:43 but last i checked pretty bad on speed 2010-09-27T23:49:00 top 3 bots are python 2010-09-27T23:49:07 lol 2010-09-27T23:49:19 5 top 10 bots are c++ though 2010-09-27T23:49:24 davidjliu, sigh : like we wondered about above, if you are ahead, maybe you can wear down the closest enemy with a constant stream. hasn't worked for me here, but.... 2010-09-27T23:50:07 dmj111: I'm more concerned about strategies for when behind... beating the enemy when ahead is not as hard :D 2010-09-27T23:50:13 dmj111, sigh: I think it worked against me here http://ai-contest.com/visualizer.php?game_id=4784428 2010-09-27T23:50:16 * dmj111 laughs... 2010-09-27T23:50:16 ha, yeah, true 2010-09-27T23:50:33 lol 2010-09-27T23:50:35 * Sylph checks 2010-09-27T23:51:26 hrm, why did you pick small planets, david 2010-09-27T23:51:31 yeah, sort of like that. and longing seems to beat everyone, at least once. 2010-09-27T23:51:37 ok, at least me. 2010-09-27T23:51:56 I think the small planets were more efficient 2010-09-27T23:52:07 * Sylph nods 2010-09-27T23:52:08 and yeah, I always have a loss to extrapostrophe 2010-09-27T23:52:11 I think position of planets is more important than growth 2010-09-27T23:52:20 ha, yeah, I just consider distance to planets 2010-09-27T23:52:25 davidjliu: same 2010-09-27T23:52:26 20 units vs 47 units, and you get 2 vs 3 growth or something ? 2010-09-27T23:52:45 and before I forget... about the playing closer rankings... things are much better than they were most of the tron competition, so I will be content if they aren't changed. 2010-09-27T23:52:50 yeah, planet size and position matters a lot 2010-09-27T23:53:07 i mean unit on the planet 2010-09-27T23:53:09 also lost to extrapostrophe... maybe multiple times. 2010-09-27T23:53:27 yeah... 2010-09-27T23:54:58 jmcarthur: i just realized you were the one from defense... hyour nick and j3camero looked the same in my small font. 2010-09-27T23:55:15 I've also been seeing the problem that a lot of games have only showed up on one of the two profiles that played the game 2010-09-27T23:55:17 smarter than me if you got out in two years. 2010-09-27T23:55:34 davidjliu: i think that is because of a resubmit. 2010-09-27T23:55:50 if you want a game, make sure to save it off... 2010-09-27T23:56:27 I mean consider http://ai-contest.com/profile.php?user_id=7935's latest game against Luksian, which is after Luksian resubmitted, but not on his profile 2010-09-27T23:57:19 ha, oops, the 's got caught in the link... 2010-09-27T23:57:21 http://ai-contest.com/profile.php?user_id=7935 2010-09-27T23:57:49 Top 10 players: dmj111(3657), adumlah(3499), felixcoto(3480), davidjliu(3479), sequoh(3463), Rexxar(3462), sighbot(3439), kcm1700(3408), Loks1982(3370), bix0r4ever(3369) 2010-09-27T23:57:51 davidjliu: good point. never noticed that before. 2010-09-27T23:58:04 it happens quite a bit, I think 2010-09-27T23:58:10 davidjliu: we both dropped a place :P 2010-09-27T23:58:23 lol 2010-09-27T23:58:24 *** bhasker has quit IRC (Read error: Connection timed out) 2010-09-27T23:58:24 ohno 2010-09-27T23:58:33 note: without having played a game in, like, hours 2010-09-27T23:58:39 lol 2010-09-27T23:58:40 yeah 2010-09-27T23:58:57 I get more game if my bot play a gagme within 5 minutes of joining 2010-09-27T23:59:02 for some reason 2010-09-27T23:59:10 yeah, they do that on purpose 2010-09-27T23:59:10 if it don't, it plays every 1-3 hours 2010-09-27T23:59:23 sleep time 2010-09-27T23:59:25 so i keep resubmitting 2010-09-27T23:59:26 see ya, everybody 2010-09-27T23:59:29 :| 2010-09-27T23:59:32 see ya. 2010-09-27T23:59:33 til it plays a game within minutes 2010-09-27T23:59:34 cya