2011-12-21T00:00:12 yeah, don't think i remember his bot doing that before 2011-12-21T00:00:44 i think i'm paying the price for not getting many early games, my catch up games are all filled with tough opponents 2011-12-21T00:01:00 but still, my bot is not living up to expectations so far 2011-12-21T00:01:09 ...and panela 2011-12-21T00:01:13 who the hell is that 2011-12-21T00:01:25 ChrisH: :( on that first part 2011-12-21T00:01:40 my bot seems to have done okay there, hopefully it evens out eventually 2011-12-21T00:02:23 I don't know who panela is, but he sent more ants my way early and that bottle me up as well 2011-12-21T00:02:41 strcat can be a verrry bad neighbor like that too 2011-12-21T00:03:12 *** Migi32 has quit IRC (Remote host closed the connection) 2011-12-21T00:03:13 but it looks like my real failure was not getting to the food store to the SE of my hill. one ant got locked in combat just north of that area and never saw the food 2011-12-21T00:03:20 i don't think you'll like me as a neighbour :P 2011-12-21T00:03:39 *** anti has joined #aichallenge 2011-12-21T00:03:51 if i had a couple ants grab that food I would have had a better chance 2011-12-21T00:03:54 anti: who are you? 2011-12-21T00:04:03 that's my name :P 2011-12-21T00:04:23 this is also anti.. 2011-12-21T00:04:33 mcstar i bet 2011-12-21T00:04:38 http://aichallenge.org/profile_games.php?user=12632&page=1 2011-12-21T00:04:48 *** praveen__ has joined #aichallenge 2011-12-21T00:04:51 this guy played 48 games.. 2011-12-21T00:05:20 how about this as a refinement to ants: 2011-12-21T00:05:25 admin, when is the cut off? 2011-12-21T00:05:32 when you are in combat+2 radius of an enemy you can exchange data 2011-12-21T00:05:41 anti: takes a while to figure out how good a starter bot is compared to other starter bots. 2011-12-21T00:05:52 say 1 bit per ant, based on whichever side has the fewest ants 2011-12-21T00:06:22 or maybe it's not one bit but values from a small enum 2011-12-21T00:06:43 You can do that with a special handshake dance, can't you, BJ? 2011-12-21T00:06:46 praveen__: it hasn't been announced yet 2011-12-21T00:07:23 Scryer: that seems like it would be a waste of turns 2011-12-21T00:07:24 Scryer: yes, but in a tight space it's not practical, and might expose you to attack 2011-12-21T00:07:34 maybe if enemies could see you swap places 2011-12-21T00:07:36 but they can't 2011-12-21T00:07:43 *** jstrong has joined #aichallenge 2011-12-21T00:07:54 ChrisH_: Ya, I am aiming at 150 rank, but am afraid I wont get enough to get to that 2011-12-21T00:07:58 Depends - if your alliance proposal is accepted you might both end up ahead. 2011-12-21T00:08:04 anyway maybe if ants get together over the course of a few turns they could negotiate something like "cease hostilities here for 20 turns" 2011-12-21T00:08:10 praveen__: where are you now? 2011-12-21T00:08:22 413 2011-12-21T00:08:28 praveen__: you're fine 2011-12-21T00:08:36 praveen__: whatever the first cut is you will make it 2011-12-21T00:09:04 ChrisH: Ya I understand that. still I feel its getting late 2011-12-21T00:09:08 unless you're 1000 places below where you were before you're probably just fine 2011-12-21T00:09:13 i think they only plan to cut less than the bottom half off 2011-12-21T00:09:17 praveen__: no no it's super early 2011-12-21T00:09:38 praveen__: games will come faster once the cuts start happening, so "time" will accellerate 2011-12-21T00:09:39 I was at about 80 the last time I got a game, 11h 22m ago, but have been gradually sliding to (now) 152. 2011-12-21T00:10:05 BenJackson: Yes I agree 2011-12-21T00:10:32 oop, xathis jumped me 2011-12-21T00:10:35 he must have gotten a game 2011-12-21T00:12:41 He's strategic gold. Nice game. 2011-12-21T00:13:56 *** anti has quit IRC (Ping timeout: 258 seconds) 2011-12-21T00:16:14 reading memetix code 2011-12-21T00:16:25 it's so radically different from anything i would've imagined, heh 2011-12-21T00:16:48 he's not a programmer 2011-12-21T00:17:05 I hope people reading MY code aren't thinking too harshly :) 2011-12-21T00:17:19 memetix is not? 2011-12-21T00:17:46 how long are we expecting the contest to run for now? 2011-12-21T00:17:49 *** CowTipperVirus has quit IRC (Quit: Page closed) 2011-12-21T00:17:51 until 22 2011-12-21T00:17:53 people were still saying under a week the other day 2011-12-21T00:18:04 jstrong: his profile (and forum post) says psychotherapist 2011-12-21T00:18:17 antimatroid: didn't they put a date on the front page? 2011-12-21T00:18:24 but it seems impractically soon 2011-12-21T00:18:28 benjackson: ah okay 2011-12-21T00:18:30 BenJackson: i think we all understand code is a mess 2011-12-21T00:18:38 i'm pretty proud of mine 2011-12-21T00:18:38 :P 2011-12-21T00:18:41 in cleanness 2011-12-21T00:18:45 i mean there are messy areas but 2011-12-21T00:19:00 Looks like meduza slipped in a solid improvement just before the deadline. 2011-12-21T00:19:03 antimatroid: but when you see how I got distracted with template metaprogramming to solve a silly problem you'll probabyl want to smack me :) 2011-12-21T00:19:09 yeah, we might want to revise the end date 2011-12-21T00:19:21 BenJackson: yeah, he could refactor stuff, but i mean the logic is just out of the box 2011-12-21T00:19:25 Scryer: I was trying to figure out if that's true or just randomness so far 2011-12-21T00:19:48 aichallenge: janzert epsilon * r3793256 / website/api_get_task.php : Try much harder to get a matchup for the worker before giving up - http://git.io/tjnFqQ 2011-12-21T00:22:10 *** Areks has joined #aichallenge 2011-12-21T00:23:27 *** xyxan has joined #aichallenge 2011-12-21T00:23:37 *** alc has joined #aichallenge 2011-12-21T00:23:57 Hello Guyz 2011-12-21T00:25:05 xathis eliminated by meduza :O 2011-12-21T00:27:46 Does that mean you're ready to place your bet? :) 2011-12-21T00:28:41 no man :P 2011-12-21T00:29:15 xathis was never eliminated :P throughout this tournament 2011-12-21T00:29:41 any ideea when signups will be available again? 2011-12-21T00:30:43 The deadline for submitting bots is past. No more new ones in the tournament. 2011-12-21T00:31:28 janzert: There must be some twist :P change game settings like viewradius & attackradius 2011-12-21T00:32:22 http://tcpants.com/player/Accoun12301 let play 2011-12-21T00:33:19 *** UncleVasya has joined #aichallenge 2011-12-21T00:33:47 I lo-o-ove turn limit: http://aichallenge.org/visualizer.php?game=299872&user=53 2011-12-21T00:34:12 *** amstan_ has joined #aichallenge 2011-12-21T00:34:12 *** ChanServ sets mode: +o amstan_ 2011-12-21T00:34:53 hhaa u are lucky 2011-12-21T00:34:57 *** rlanday has joined #aichallenge 2011-12-21T00:34:59 *** gaussin has quit IRC (Quit: Leaving.) 2011-12-21T00:35:51 UncleVasya: also lots of beneficiaries of "rank stabilized" when one guy doesn't lose his hill and comes in "2nd" 2011-12-21T00:36:12 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-12-21T00:37:10 1-1 games are good :P 2011-12-21T00:37:32 games with more Botz sometimes seems biased 2011-12-21T00:37:35 UncleVasya: wow, at the beginning I thought you were going to avoid being taken by itzkow but not win. congrats 2011-12-21T00:38:19 :) 2011-12-21T00:38:51 *** Zedenstein has quit IRC (Ping timeout: 258 seconds) 2011-12-21T00:40:47 heheh, near the same: http://aichallenge.org/visualizer.php?game=298334&user=53 2011-12-21T00:41:28 :P u seems lucky 2011-12-21T00:41:43 yep 2011-12-21T00:42:06 *** praveen__ has quit IRC (Quit: Page closed) 2011-12-21T00:42:44 :( my no game since last 8 he 13 min :( still it saying 600 mins remaining 2011-12-21T00:43:01 *** jstrong has quit IRC (Read error: Connection reset by peer) 2011-12-21T00:43:09 *** jstrong has joined #aichallenge 2011-12-21T00:43:30 how many do you have already? 2011-12-21T00:43:34 xyxan: they started prioritizing people with few games 2011-12-21T00:43:38 xyxan: with the change in scheduling to work through all the lowest game count bots first, that counter now behaves different 2011-12-21T00:43:40 15 :P 2011-12-21T00:44:03 xyxan: so it won't go down much until all the bots with fewer games than you have caught up 2011-12-21T00:44:58 xyxan: and once it starts going down a little faster that means you could get a game at any moment because it doesn't do a good job of knowing where you are within the group with the same number of games as yu 2011-12-21T00:45:02 ahan right :) but guyz there must be some change in game settings like viewradius & attackradius 2011-12-21T00:45:10 xyxan:,so I am luck again to have 11 games and to play few times more befor you will do :p 2011-12-21T00:45:18 *lucky 2011-12-21T00:45:27 *before 2011-12-21T00:45:32 yup :P 2011-12-21T00:45:44 hmm, always forget to type last literas... 2011-12-21T00:46:23 hummm, btw its a great experience :D 2011-12-21T00:47:46 any ideas about next competitoin? 2011-12-21T00:48:04 There are suggestions on the forum. 2011-12-21T00:49:14 infact i started late this game, but still i managed to reach top 800 :D 2011-12-21T00:49:24 i think the engine is working its way through the 11 game bots now 2011-12-21T00:49:49 yeah people with fewer games must be given priority 2011-12-21T00:49:51 *** thagomizr has joined #aichallenge 2011-12-21T00:50:39 There was some talk about the first cut being after 10 games, but I guess that was unofficial conjecture amongst the players rather than the organizers. 2011-12-21T00:51:12 Scryer: a little more than that, that was my goal for the minimum number of games before starting the cutoff 2011-12-21T00:51:23 in my opinion, starter botz must not participate in final tournament 2011-12-21T00:51:34 Scryer: I think the organizers were talking about that number too, but I haven't seen any chatter here recently about them implementing it yet 2011-12-21T00:52:02 because some people have gained a lot of benifit because of this 2011-12-21T00:52:07 ahh, janzert is here to give more official answers, good 2011-12-21T00:52:31 janzert: what about change in game settings :( 2011-12-21T00:52:37 *** rlanday_ has joined #aichallenge 2011-12-21T00:52:46 as a continuation tournament of some sort? 2011-12-21T00:53:02 *** rlanday has quit IRC (Ping timeout: 258 seconds) 2011-12-21T00:53:04 I don't think we have money or volunteer willpower for it 2011-12-21T00:53:16 How about increasing vis and combat radius after the first cutoff. :) (Joke, joke!!) 2011-12-21T00:53:17 we're ready for a break here :} 2011-12-21T00:53:30 I'm still just hanging in the top 1000, probably won't make top 800 ... there's definitely a bug where it doesn't grab holes even when it has the chance, especially if a food pops up near the hole to distract the ant. I didn't get the tuning quite right on that one. 2011-12-21T00:53:47 janzert: increase view and attack radius :D 2011-12-21T00:54:36 That's bitten me about 5 times where they dawdled around not grabbing holes, and probably that's at least one "skills" point right there 2011-12-21T00:54:58 *** Paradoxial has quit IRC (Quit: :) 2011-12-21T00:55:33 lol and my BOT is always surrounded by top ranked botz so i cant go for food leaving my hill :D 2011-12-21T00:56:11 *** kost_ has joined #aichallenge 2011-12-21T00:59:53 how many games will be played till first CutOff? 2011-12-21T00:59:54 Wow, 26 games and he's still getting 'em! http://aichallenge.org/profile.php?user=9606 ... ImportANT is well-named! 2011-12-21T01:00:16 at least 11 ;) 2011-12-21T01:01:00 ImportAnt is doing excellent in reverse direction :D 2011-12-21T01:01:40 janzert: nice tautology 2011-12-21T01:02:18 *** gcflymoto has joined #aichallenge 2011-12-21T01:02:57 how many top botz will play after first CutOff? 2011-12-21T01:04:05 *** Zaphus has joined #aichallenge 2011-12-21T01:04:51 *** antimatroid has quit IRC (Quit: Page closed) 2011-12-21T01:06:04 Zaphus: hi. Finally I've got enough games to come to the first page. 11 now. 2011-12-21T01:06:45 UncleVasya: Congrats! unfortunately I've lost the last 4 or 5 of my games, so I won't be joining you in the 100 :-( 2011-12-21T01:06:46 BenJackson at top now :D 2011-12-21T01:07:00 Hey, I got to the #1 spot in 4 games! I have the screenshot to prove it. :) 2011-12-21T01:07:53 I am at 730 :P and no game since last 9 hrz 2011-12-21T01:07:55 when you'll come to the first place after xathis played 20 -- then give me a screenshot :D 2011-12-21T01:08:05 lolx 2011-12-21T01:08:48 I wish to have a 1-1 multihill game with xathis :P 2011-12-21T01:08:49 xyxan: yeah, you have 15 games, you need to wait for others to catch up :-) 2011-12-21T01:09:04 My last game was 11.5 hours ago - it'll be a while. 2011-12-21T01:09:13 http://forums.aichallenge.org/viewtopic.php?f=27&t=2183&p=13656#p13656 2011-12-21T01:09:16 O.o 2011-12-21T01:09:48 there's some information on the cutoff plans 2011-12-21T01:10:30 good post janzert, thanks 2011-12-21T01:10:39 Thanks, janzert. 2011-12-21T01:10:48 Janzert - great work 2011-12-21T01:11:03 hey, how did BenJackson end up at #1 2011-12-21T01:11:09 hmm, what'd I break now ;) 2011-12-21T01:11:10 cutoff around 500 :O It should be 1000 :( 2011-12-21T01:11:12 the new scheduling seems to be working well (at least for those who were behind... probably doesnt feel as good for xyxan ) 2011-12-21T01:11:34 Pretty sure that means the final cutoff, xyxan, not the initial cutoff. 2011-12-21T01:11:37 Its good Zaphus :D 2011-12-21T01:11:51 *** anti has joined #aichallenge 2011-12-21T01:11:53 oh gr8 :D 2011-12-21T01:12:11 *** Areks has quit IRC (Ping timeout: 268 seconds) 2011-12-21T01:12:11 The starter bots are somewhere around 2-3000, I think. 2011-12-21T01:12:17 yes, that is the farthest I expect it to go 2011-12-21T01:12:42 greedybot is at 2080 or so, that might be your marker 2011-12-21T01:13:10 yeah :D 2011-12-21T01:13:38 we're down to 586 bots with 11 games 2011-12-21T01:16:38 ChrisH: do you also know the max? I saw one in the 40s as I was glancing around by hand. 2011-12-21T01:17:14 McLeopold did a query earlier that showed a bot with 66 games 2011-12-21T01:17:22 that was the max 2011-12-21T01:17:34 Ah. Thanks. 2011-12-21T01:17:43 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-12-21T01:17:49 *** ltriant has quit IRC (Quit: Computer has gone to sleep) 2011-12-21T01:17:49 but the majority of bots are now clusting just above the minimum in the 11-14 range I believe 2011-12-21T01:18:01 http://pastebin.com/rxCe0vTY 2011-12-21T01:18:08 66 :O lolx crazy 2011-12-21T01:18:40 yeah, I think it's a starter bot, the ultimate weed out test bot 2011-12-21T01:18:48 Thanks, janzert. Looks like the bottom is clumping pretty well. 2011-12-21T01:19:51 *** Zaphus has quit IRC (Quit: Page closed) 2011-12-21T01:22:10 *** rlanday_ has quit IRC (Ping timeout: 258 seconds) 2011-12-21T01:26:01 *** UncleVasya has quit IRC (Read error: Connection reset by peer) 2011-12-21T01:26:27 *** UncleVasya has joined #aichallenge 2011-12-21T01:28:43 *** Scryer has quit IRC (Remote host closed the connection) 2011-12-21T01:30:13 *** anti has quit IRC (Ping timeout: 258 seconds) 2011-12-21T01:37:19 *** bucash has joined #aichallenge 2011-12-21T01:38:38 *** bucash has quit IRC (Client Quit) 2011-12-21T01:39:33 yes, please change the combat radius now 2011-12-21T01:39:41 Unlucky: http://aichallenge.org/visualizer.php?game=300765 2011-12-21T01:39:57 watch this game and see how ranking system works :d 2011-12-21T01:40:11 heh oops 2011-12-21T01:40:20 *** Areks has joined #aichallenge 2011-12-21T01:40:38 starter Bot win after losing whatever he had :D 2011-12-21T01:41:28 @rankings 2011-12-21T01:41:29 BenJackson: Top 10 players: xathis(82.5), BenJackson(81.7), meduza(81.4), FlagCapper(81.4), delineate(80.0), runevision(79.8), lazarant(79.8), _Atsky_(79.6), teapotahedron(79.6), Speedy_Consoles(79.4) 2011-12-21T01:41:39 oop missed my second shot at #1 apparently 2011-12-21T01:42:08 lolx xathis finally on top :D 2011-12-21T01:42:20 that must have just happened 2011-12-21T01:42:30 you were #1 a few seconds ago ben 2011-12-21T01:42:55 a1kon but i still agree with you about changing combat radius & view radius 2011-12-21T01:43:10 my bot would cope with that just fine :) 2011-12-21T01:44:18 haha i have another idea :P @janzert what about map with no water :D 2011-12-21T01:44:26 boring 2011-12-21T01:44:47 an open battle ground will be a good measure :D 2011-12-21T01:45:05 wow that was a nice battle in my last game 2011-12-21T01:45:14 I broke out of my slump: http://aichallenge.org/visualizer.php?game=300785&user=10 2011-12-21T01:45:18 knowing I didn't have to defend my interior really helped 2011-12-21T01:45:33 *** meduza has joined #aichallenge 2011-12-21T01:45:43 and I love that moment where my line at blue fails (he was good at getting on the outside of those lines!) a huge hunk of my ants dove over there 2011-12-21T01:45:48 *** jstrong has quit IRC (Quit: Leaving.) 2011-12-21T01:46:35 I'm not even entirely sure how the rebalance at about turn 400 happened but I'm not sure I've ever seen anything like it 2011-12-21T01:46:45 *** Randex has quit IRC (Quit: Page closed) 2011-12-21T01:48:24 *** erdman has joined #aichallenge 2011-12-21T01:49:15 BenJackson: it looks like blue might be a few steps closer to you hill at that point than yellow 2011-12-21T01:51:20 the mu scores for the bots in the 16-60 range are pretty tightly packed 2011-12-21T01:51:28 a lot of sorting out to do there yet 2011-12-21T01:51:37 *** woot_ has joined #aichallenge 2011-12-21T01:51:42 just saw something i wanted to bring to attention of organizers and to those otherwise in the know 2011-12-21T01:51:52 so i'm accustomed to seeing players get a flurry of games 2011-12-21T01:52:04 but this player got a flurry of games all on same map 2011-12-21T01:52:17 http://aichallenge.org/profile.php?user=2321 2011-12-21T01:52:27 *** Palmik has joined #aichallenge 2011-12-21T01:52:28 which seems quite odd 2011-12-21T01:52:46 yes, 2011-12-21T01:52:47 :O 2011-12-21T01:53:06 is that to be expcted with the landmarking thing? 2011-12-21T01:53:07 really strange 2011-12-21T01:53:10 I believe McLeopold said he thinks there is still a race condiditon in the scheduler somewhere, and that profile seems to be evidence of that 2011-12-21T01:53:35 ok, good, i'm glad it's viewed as a problem 2011-12-21T01:53:38 erdman: no, i think that has more to do with a bug in the timing of when matches are created and when the bot gets removed from the queue 2011-12-21T01:54:35 chrish: thanks, jsut wanted to make sure u guys are aware 2011-12-21T01:54:50 janzert, McLeopold here is another profile that shows a possible race condidtion in scheduling: http://aichallenge.org/profile.php?user=2321 2011-12-21T01:55:15 thanks erdman 2011-12-21T01:58:16 Yes, that should be avoided by the change to stop the worker request from generating a matchup and instead make it wait for one to become available 2011-12-21T01:58:28 that change was made earlier today 2011-12-21T01:59:48 I got a row of games on the same map and I thought it was a helpful feature. Very interesting comparison to see different opponents on the same map. 2011-12-21T02:00:03 *** woot_ has quit IRC (Quit: Page closed) 2011-12-21T02:01:48 *** valydo has joined #aichallenge 2011-12-21T02:05:13 *** TL__ has joined #aichallenge 2011-12-21T02:05:23 *** raulis has joined #aichallenge 2011-12-21T02:07:47 *** UncleVasya has quit IRC (Ping timeout: 252 seconds) 2011-12-21T02:07:56 *** raulis has quit IRC (Client Quit) 2011-12-21T02:09:48 hmm, Liquid was in top 100 shortly before open submissions closed and then submitted a starter 3 hours before submissions closed :( http://aichallenge.org/profile.php?user=11105 2011-12-21T02:10:20 got frustrated? 2011-12-21T02:11:59 he may have another account :O 2011-12-21T02:13:17 another way to remove starter botz is to check similarity between different botz code :D 2011-12-21T02:13:32 *** gcflymoto has quit IRC (Ping timeout: 258 seconds) 2011-12-21T02:14:19 *** Westicles has joined #aichallenge 2011-12-21T02:16:28 *** bmh has joined #aichallenge 2011-12-21T02:17:48 hi. have we applied the cutoff already? 2011-12-21T02:18:01 not yet 2011-12-21T02:19:22 *** smjms has quit IRC (Quit: Konversation terminated!) 2011-12-21T02:19:28 *** muMeson has joined #aichallenge 2011-12-21T02:19:38 *** smjms has joined #aichallenge 2011-12-21T02:22:21 *** meduza has quit IRC (Ping timeout: 258 seconds) 2011-12-21T02:24:55 *** Ashoka has joined #aichallenge 2011-12-21T02:29:36 After first Cutoff, will all ranking and skill start from zero?? 2011-12-21T02:29:52 no 2011-12-21T02:30:19 then people having high mu will get benifit :D 2011-12-21T02:30:19 the cutoff doesn't effect the ranking at all just the pairing 2011-12-21T02:32:56 *** valydo has quit IRC (Ping timeout: 240 seconds) 2011-12-21T02:34:10 *** UncleVasya has joined #aichallenge 2011-12-21T02:34:27 does anyone know the right syntax for making a numbered list in the forums? 2011-12-21T02:34:43 I can only get bulletted lists to work 2011-12-21T02:34:48 *** bmh has left #aichallenge 2011-12-21T02:34:55 # instead of *? 2011-12-21T02:35:18 i'll try that 2011-12-21T02:35:49 nope 2011-12-21T02:36:01 ChrisH_: back at a real computer and watched your win 2011-12-21T02:36:10 It's not immediately obvious if the hill thing helped 2011-12-21T02:36:12 did you analyzet hat? 2011-12-21T02:36:44 *** smjms has quit IRC (Quit: Konversation terminated!) 2011-12-21T02:37:24 i only lost one hill the whole game because i could put defenders on the spot when needed 2011-12-21T02:38:20 *** UncleVasya has quit IRC (Remote host closed the connection) 2011-12-21T02:38:22 and the hill at 115, 64 is never really threatened so it spawns fewer ants for most of the game 2011-12-21T02:39:16 the hill at 138, 0 seems to spawn a lot of ants that help defend ant then take several nearby hills 2011-12-21T02:39:41 that's what I saw anyway 2011-12-21T02:40:05 from forum:> One of my friends who essentially used the starter bot and just added pathfinding (for each pair of ants-food like in the tutorial) and timeout handling and no more sophisticated stuff is around rank 1500 already. 2011-12-21T02:40:13 that's better than greedy I think 2011-12-21T02:40:20 of course greedy was timing out like crazy in the finals 2011-12-21T02:40:57 wow the non-native english speakers in the forum are misreading janzert 2011-12-21T02:44:00 let me try some bold, large, colorful text 2011-12-21T02:45:16 Added a clarification edit as well 2011-12-21T02:45:26 well shit you'd better edit my quoting/bolding of you :) 2011-12-21T02:45:53 nah, the edit is likely to be missed 2011-12-21T02:46:07 hopefully your won't ;) 2011-12-21T02:46:16 yours 2011-12-21T02:49:15 *** antimatroid has joined #aichallenge 2011-12-21T02:50:03 http://aichallenge.org/visualizer.php?game=300798&user=98 goddamnit 2011-12-21T02:50:03 I just posted the first installment of my bot description on the forums. Mostly a high level overview, but its a start. 2011-12-21T02:51:49 xathis is in place 2011-12-21T02:52:16 I was playing skyrim when I jumped him briefly :( 2011-12-21T02:52:39 even though it will make no ultimate difference it's entertaining riding out this phase in 1st/2nd rather than, you know, 85th or something 2011-12-21T02:53:07 BenJackson: you probably get a small advantage actually :) 2011-12-21T02:53:41 BenJackson: you need to correct the typo of "note a starter bot" in your post 2011-12-21T02:53:46 I am expecting a phse where all my games are vs flag and xathis and lazarant where I'm whipped down a few notches 2011-12-21T02:54:04 followed by a period where I beat on some top 20-30 bots as my sigma drops 2011-12-21T02:54:28 my bot dropped lots in rank recently :( 2011-12-21T02:54:53 ChrisH_: danke 2011-12-21T02:55:29 ChrisH_: thanks for the post 2011-12-21T02:56:18 *** Raulis has joined #aichallenge 2011-12-21T02:56:51 so, as i said a bit ago, i have posted an overview of my bot: http://forums.aichallenge.org/viewtopic.php?f=24&t=2202 I would love to hear comments on style or organization, etc. so I can adjust as I add more. 2011-12-21T02:57:01 reading it now 2011-12-21T02:57:23 I was pondering how Dijkstra w/cost was different from bfs w/cost in this case 2011-12-21T02:57:27 but I assume you will elucidate 2011-12-21T02:58:01 antimatroid : you deserved little more than 2 points in that game. u survived till end and had a decent army as well... may be your bot should be taking few risks and attack more. 2011-12-21T02:58:17 BenJackson: I don't think a BFS with costs is a BFS anymore 2011-12-21T02:58:21 BenJackson: they are the same AFAIK 2011-12-21T02:58:22 *** Kettling has left #aichallenge ("Leaving") 2011-12-21T02:58:53 I guess with costs the bfs revisits squares, dijkstra would not 2011-12-21T02:59:30 @rankings 2011-12-21T02:59:37 Ashoka: Top 10 players: xathis(82.5), BenJackson(81.7), meduza(81.4), FlagCapper(81.4), delineate(80.0), runevision(79.8), lazarant(79.8), _Atsky_(79.6), teapotahedron(79.6), Speedy_Consoles(79.4) 2011-12-21T02:59:57 dijkstra's does not revisit, i've always thought of dijkstra's as a BFS that allows for possitive weights other than 1 2011-12-21T03:00:05 Just curious, how many games do you think top bots will have when this is over/ 2011-12-21T03:00:19 one of the organizers said 200ish 2011-12-21T03:00:27 but they've also said the 22nd which seems incompatible with that 2011-12-21T03:00:51 BenJackson: Thanks. Any reason why it should be a few thousand? 2011-12-21T03:00:57 shouldn't 2011-12-21T03:01:06 time? money? 2011-12-21T03:01:22 janzert was talking about a final cutoff around 500 2011-12-21T03:01:41 even assuming 5p games average 100,000 games is a lot on the scale of the contest 2011-12-21T03:01:48 Oh. I was thinking tighter cutoffs 2011-12-21T03:01:54 only 300,000 have been played total so far 2011-12-21T03:02:14 *** Raulis has quit IRC (Quit: Page closed) 2011-12-21T03:02:23 I suspect if there's some great unsettled duel in the top N there might be a further cutoff to straighten it out 2011-12-21T03:02:40 *** raulis has joined #aichallenge 2011-12-21T03:02:48 although really these things need to be decided well in advance to avoid changing the outcome by late decision 2011-12-21T03:02:52 janzert: ^ 2011-12-21T03:03:21 janzert: ^ ^ 2011-12-21T03:03:57 *** epicmonkey has joined #aichallenge 2011-12-21T03:04:16 200 was a hoped for guess that will probably not play out now that the rate we can play games is so much lower than expected 2011-12-21T03:04:46 so the time cutoff is fixed? 2011-12-21T03:04:53 there are problems with making the cutoff too tight so 250 is basically it 2011-12-21T03:05:06 the end of competition isn't a hard fix so may be pushed out some 2011-12-21T03:05:15 there has been a bit of talk about doing that 2011-12-21T03:05:43 but probably only a day 2011-12-21T03:06:31 should the "it only takes 5 minutes to submit..." paragraph be removed from the front page? 2011-12-21T03:06:47 lol ^ 2011-12-21T03:07:09 probably :) 2011-12-21T03:07:40 well, I don't know 2011-12-21T03:07:57 the ai-class book market is softening up 2011-12-21T03:07:59 anyway the big heading at the top should tip you off I think 2011-12-21T03:08:11 *** mj41 has joined #aichallenge 2011-12-21T03:08:19 anyone from ml-class? 2011-12-21T03:08:21 *** TL__ has quit IRC (Ping timeout: 258 seconds) 2011-12-21T03:08:32 janzert: At this point, the game rate is purely a function of CPU power/time, right? Nothing that can be helped with? 2011-12-21T03:08:57 or the matchmaker 2011-12-21T03:09:18 janzert: add a web form to suggest matchups. that will keep things going 2011-12-21T03:09:58 the game rate is limited by the database performance actually 2011-12-21T03:09:58 *** Jak_o_Shadows has quit IRC (Remote host closed the connection) 2011-12-21T03:10:06 inserting games and creating pairings 2011-12-21T03:10:24 we tried not inserting games 2011-12-21T03:10:31 :) 2011-12-21T03:10:38 250 seems tight for me and when will this be in action> 2011-12-21T03:10:40 ? 2011-12-21T03:11:24 Is there scope to help with that? Is it review-query-to-try-to-improve slow, or just hardware-limitation slow? 2011-12-21T03:11:30 but all people above cutoff must have equal number of games :D 2011-12-21T03:11:31 *** bob__ has joined #aichallenge 2011-12-21T03:13:12 I'm not much of a database/sql guy, that's McLeopold's expertise 2011-12-21T03:13:13 *** Ashoka has quit IRC (Quit: Page closed) 2011-12-21T03:13:36 also I'd rather not make much change to it in the middle of finals if possible 2011-12-21T03:14:27 Fair enough 2011-12-21T03:14:38 *** epicmonkey has quit IRC (Ping timeout: 240 seconds) 2011-12-21T03:14:55 *** moduspwned has quit IRC (Quit: WeeChat 0.3.6) 2011-12-21T03:15:39 if it's a simple missing index and someone can tell me it's low risk to have the database fall over from adding it I'd be willing to do it though ;) 2011-12-21T03:15:45 *** smiley1983 has joined #aichallenge 2011-12-21T03:15:56 but my understanding is that McLeopold has gone through that sort of thing 2011-12-21T03:16:01 *** bob__ has quit IRC (Ping timeout: 258 seconds) 2011-12-21T03:16:14 *** muMeson is now known as hadron 2011-12-21T03:19:31 *** mojoli has joined #aichallenge 2011-12-21T03:20:13 *** ChrisH_ has quit IRC (Quit: Page closed) 2011-12-21T03:21:39 *** codetiger has joined #aichallenge 2011-12-21T03:22:07 Any ETA when pairing cut off is going to be set... 2011-12-21T03:22:40 with expected bots on top, I am happy the contest is in the right direction 2011-12-21T03:23:37 @rankings 2011-12-21T03:23:42 BenJackson: Top 10 players: xathis(82.5), BenJackson(81.7), meduza(81.4), FlagCapper(81.4), delineate(80.0), runevision(79.8), lazarant(79.8), _Atsky_(79.6), teapotahedron(79.6), Speedy_Consoles(79.4) 2011-12-21T03:23:46 I agree 2011-12-21T03:23:50 carry on 2011-12-21T03:24:34 *** mattswe has quit IRC (Remote host closed the connection) 2011-12-21T03:26:08 BenJackson: seems you took down xathis in one of your games together with other top bots 2011-12-21T03:26:37 did I even encounter him in that game? 2011-12-21T03:27:05 oh, I guess so, forgot that one 2011-12-21T03:27:06 no, but just his bad placing... 2011-12-21T03:27:45 just watching the game 2011-12-21T03:28:11 *** mansfeld has quit IRC (Quit: Changing server) 2011-12-21T03:28:13 mostly I participated in a ginormous stalemate 2011-12-21T03:28:43 I probably would have done better if I'd lost one hill early 2011-12-21T03:29:52 *** Zaphus has joined #aichallenge 2011-12-21T03:29:57 *** mattswe has joined #aichallenge 2011-12-21T03:30:16 *** Blkt has joined #aichallenge 2011-12-21T03:31:16 *** Zeta-Hoernchen has joined #aichallenge 2011-12-21T03:31:25 http://aichallenge.org/visualizer.php?game=298375&user=7935 2011-12-21T03:32:29 oh cool 2011-12-21T03:32:36 I've been waiting to see the swastica map in action 2011-12-21T03:32:38 forgot to search it 2011-12-21T03:32:44 ChrisH_: ^ 2011-12-21T03:33:00 *** xathis has joined #aichallenge 2011-12-21T03:33:33 here comes xathis :P 2011-12-21T03:34:20 Zeta-Hoernchen : I'd love to see a version of the visualiser that showed 'ownership' of the map - places where the different bots own the territory, that match would be interesting with that view 2011-12-21T03:34:36 *** raulis has quit IRC (Quit: Page closed) 2011-12-21T03:34:49 *** Zeta-Hoernchen is now known as Speedy_Consoles 2011-12-21T03:35:04 good morning everyone 2011-12-21T03:35:37 (yes, it is a /amsg, but I love to greet everyone) 2011-12-21T03:35:39 http://aichallenge.org/map.php?map=maze/maze_p04_42.map 2011-12-21T03:35:46 *** bob__ has joined #aichallenge 2011-12-21T03:36:00 Blkt: evening for some :-) but g'day to you too 2011-12-21T03:36:18 :D 2011-12-21T03:36:46 kinda afternoon in Pakistan :) 2011-12-21T03:37:07 BenJackson: swastika hey... accident or someone trying to be clever ? 2011-12-21T03:37:14 Speedy_Consoles: your game might be the only one on that with a good matchup 2011-12-21T03:37:20 Zaphus: random 2011-12-21T03:37:33 not that unlikely on a 4-way symmetry map 2011-12-21T03:37:50 BenJackson fair enough, it's bound to happen with enough maps and that sort of way of creating them 2011-12-21T03:38:16 I was joking that xathis wouldn't be allowed to play it under German law 2011-12-21T03:39:52 *** TL__ has joined #aichallenge 2011-12-21T03:40:10 *** mojoli has quit IRC (Ping timeout: 258 seconds) 2011-12-21T03:41:39 http://tcpants.com/player/Accoun12301 let play 2011-12-21T03:42:34 *** bob__ has quit IRC (Quit: Page closed) 2011-12-21T03:43:24 *** codetiger has quit IRC (Remote host closed the connection) 2011-12-21T03:47:19 I need a win, I came last 4 games in a row :-( still in the top 200 but that really hurt! 2011-12-21T03:48:27 where were you before the finals? 2011-12-21T03:49:09 *** Speedy_Consoles has quit IRC (Quit: KVIrc 4.0.1 Insomnia http://www.kvirc.net/) 2011-12-21T03:50:04 similar position, around the 60-80 mark when it did well, but the increase in 2-player maps hurts my bot 2011-12-21T03:52:35 plus I made a change in the last 36 hours... it looked good on paper and in local testing, but.... you never know :-) 2011-12-21T03:53:43 *** Manwe56 has quit IRC (Quit: Page closed) 2011-12-21T03:54:16 *** Westicles has quit IRC (Quit: Page closed) 2011-12-21T03:55:06 *** Nihil688 has joined #aichallenge 2011-12-21T03:55:31 Zaphus: I've also had a string of losses, really hoping the game rate increases dramatically :) 2011-12-21T03:56:19 maaan... My ants haven't had a game in ages... The system keeps counting down, reseting and back to 500 minutes of waiting again... 2011-12-21T03:56:56 I dropped 400 ranks just cause of inactivity... 2011-12-21T03:57:47 *** Nihil688 has quit IRC (Client Quit) 2011-12-21T04:00:44 nihil688 How many games have you had ? 2011-12-21T04:05:51 *** antimatroid has quit IRC (Ping timeout: 258 seconds) 2011-12-21T04:08:08 *** meduza has joined #aichallenge 2011-12-21T04:16:07 *** erdman has quit IRC (Quit: Page closed) 2011-12-21T04:16:35 *** xathis has quit IRC (Ping timeout: 258 seconds) 2011-12-21T04:16:47 doh, got another game, lost it... 5 losses in a row :( 2011-12-21T04:17:11 *** bob__ has joined #aichallenge 2011-12-21T04:17:21 *** meduza has quit IRC (Ping timeout: 258 seconds) 2011-12-21T04:19:36 :( dropped from 611 to 749 in last 13 hrs :P my no game since last 13 hrz 2011-12-21T04:20:17 *** Jak_o_Shadows has joined #aichallenge 2011-12-21T04:20:44 *** Belerafon_L has joined #aichallenge 2011-12-21T04:25:00 *** replore_ has quit IRC (Remote host closed the connection) 2011-12-21T04:26:11 *** Surya has joined #aichallenge 2011-12-21T04:27:34 *** flowenol has joined #aichallenge 2011-12-21T04:28:54 *** smiley1983 has quit IRC (Ping timeout: 268 seconds) 2011-12-21T04:29:05 *** smiley1983 has joined #aichallenge 2011-12-21T04:30:18 *** Zaphus has quit IRC (Quit: Page closed) 2011-12-21T04:33:41 *** nha has joined #aichallenge 2011-12-21T04:35:17 Hi everyone. Did the cutoff happen during my night? :) 2011-12-21T04:35:50 nope ;) 2011-12-21T04:35:55 probably soon though 2011-12-21T04:36:10 you can see the current cutoff on the server stat page btw 2011-12-21T04:37:25 wow 28 servers up :) 2011-12-21T04:37:34 Janzert: oh ok nice :) 2011-12-21T04:37:46 you mean 30 2011-12-21T04:37:49 err 29 2011-12-21T04:37:57 30 is coming up though 2011-12-21T04:38:01 7*4 + 1 :D 2011-12-21T04:38:12 9 * 3 + 1 2011-12-21T04:39:08 oops, I'm forgetting how many I have going 2011-12-21T04:39:10 :/ 2011-12-21T04:39:24 :) 2011-12-21T04:39:32 Janzert : How many bots are you going to cutoff btw? (just an approximation) 2011-12-21T04:39:58 it will initially go to 5000 over a period of a few hours 2011-12-21T04:40:16 Ok, nice. Thx for those info 2011-12-21T04:40:21 well maybe quicker for this first one 2011-12-21T04:40:35 but 5000 is the first stopping point in any case 2011-12-21T04:43:21 janzert: soon? 2011-12-21T04:43:43 after everyone has 13 games and I do a few checks 2011-12-21T04:43:55 any estimated time? 2011-12-21T04:44:05 (sorry for being annoying but im way too stressed) 2011-12-21T04:44:11 probably an hour 2011-12-21T04:44:39 okay 2011-12-21T04:44:41 you're no where near close to it :) 2011-12-21T04:44:48 its gonna be a long wait 2011-12-21T04:44:51 *** valydo has joined #aichallenge 2011-12-21T04:45:01 janzert: well i keep asking when the bots will be cut 2011-12-21T04:45:10 i think thats annoying enough 2011-12-21T04:45:26 I mean you're no where near close to being in this cutoff 2011-12-21T04:45:35 ah´ 2011-12-21T04:45:37 no comment on being annoying :P 2011-12-21T04:45:37 i know that 2011-12-21T04:45:54 but the faster the cutoff, the faster my bot will play more 2011-12-21T04:45:59 so the more it may rise 2011-12-21T04:46:14 and im expecting it to go up a decent value in skill 2011-12-21T04:46:44 don't expect it to translate into ranks though 2011-12-21T04:47:02 will translate into country rank 2011-12-21T04:47:04 hopefully 2011-12-21T04:47:05 since everyone around you is likely to have there skill going up as well 2011-12-21T04:47:29 since i had a nice advantage over the second one in my country 2011-12-21T04:47:38 ahh, that looks to be close for you :) 2011-12-21T04:47:55 i gave up on top 100 a long time ago 2011-12-21T04:48:05 first in my country will suffice 2011-12-21T04:48:20 i was there before the cmpetition started so i have hope 2011-12-21T04:49:14 :) 2011-12-21T04:50:15 ah just do the cutoff 2011-12-21T04:50:25 * flowenol is beconing insane with this 2011-12-21T04:50:41 lol 2011-12-21T04:51:07 cutoff is not as important as change of viewradius & attackradius :D 2011-12-21T04:51:14 im too addicted to this 2011-12-21T04:51:26 xyxan: mine are not hardcoded 2011-12-21T04:51:30 my bot should be fine 2011-12-21T04:51:40 http://tcpants.com/player/Accoun12301 let play 2011-12-21T04:51:45 lots of botz will have poor performance 2011-12-21T04:52:10 flowenol :) mine is also not hardcoded 2011-12-21T04:52:25 it may have a little lower performance on combat 2011-12-21T04:52:35 if attackradius goes up 2011-12-21T04:52:41 depends on how much 2011-12-21T04:53:29 xyxan: some friends made some of those hardcoded, they may have some trouble 2011-12-21T04:53:33 attackradius2 = 7 & viewradius2 = 99 :) 2011-12-21T04:53:44 ^ will be a good choice 2011-12-21T04:54:30 not a big change 2011-12-21T04:54:38 attackradius2 = 9 & viewradius2 = 99 :) 2011-12-21T04:54:50 9 will be gud :D 2011-12-21T04:54:53 now that may affect my performance a little 2011-12-21T04:55:01 but lets see 2011-12-21T04:55:11 im not too worried with that 2011-12-21T04:55:19 I'd rather not see a change in those parameters :) 2011-12-21T04:55:22 not even me :) 2011-12-21T04:55:28 My bot would go crazy >< 2011-12-21T04:55:42 Surya: why 2011-12-21T04:56:29 Because my fighting method has been made up at the end, And I use attackRadius + some hardcoded position. Needless to say I'm not so proud of it. 2011-12-21T04:56:48 also i would suggest adding maps with no water :D <<< that will make many botz CRAZY :) 2011-12-21T04:57:00 janzert ^ ^ 2011-12-21T04:57:06 *** Manwe56 has joined #aichallenge 2011-12-21T04:57:42 Are they going to be changed? I'm scared now. 2011-12-21T04:57:50 now that i think about it, larger view radius may be better for me 2011-12-21T04:58:07 i dont think so :) there is very less probability of change in parameters 2011-12-21T04:58:46 waht if viewradius = attackradius :D 2011-12-21T04:59:14 Ok, that's what I saw on the forum. Next challenge I won't make the mistake though 2011-12-21T04:59:16 everything will go crazy :D 2011-12-21T05:02:21 xyxan: quit trying to scare people :P 2011-12-21T05:03:05 at this point no more maps will be added, no game setting changes will be made 2011-12-21T05:03:13 :O Oops 2011-12-21T05:04:04 xyxan: haha 2011-12-21T05:04:32 lmao :P whats so horrible :d 2011-12-21T05:05:10 yay playing a game 2011-12-21T05:05:32 after 12h 2011-12-21T05:06:18 If i would be one of organizer, I would have done this as soon as final tournament started :D 2011-12-21T05:07:03 flowenol : i had no game since last 12 hr 43 min 2011-12-21T05:07:16 eeck 2011-12-21T05:08:47 *** mayanks43 has joined #aichallenge 2011-12-21T05:14:51 xyxan: That's why you're not one of the organizers maybe ;) 2011-12-21T05:15:11 lmao :D 2011-12-21T05:15:27 You'd have half of the players trying to find you IRL :D 2011-12-21T05:15:44 Since everybody goes crazy with this challenge :) 2011-12-21T05:15:46 IRL? 2011-12-21T05:16:02 in real life. Geek spotted :O 2011-12-21T05:16:08 hahahahahahah 2011-12-21T05:16:34 I would be most wanted :D thats so awesome :d 2011-12-21T05:17:02 :D 2011-12-21T05:17:23 that means in order to be most wanted one must be organizer :D 2011-12-21T05:17:35 I love those 30 servers make games & games. 2011-12-21T05:17:45 making* 2011-12-21T05:17:50 yeah now servers are performing good :D 2011-12-21T05:18:34 (I'm only saying that coz Janzert said yesterday that people only complain ;)) 2011-12-21T05:18:55 what? I hope I didn't say that 2011-12-21T05:19:06 we are here to appriciate organizers countless efforts :) 2011-12-21T05:19:06 No, I'm extrapolating, don't worry 2011-12-21T05:19:12 If I did I blame it on a sleep deprived stupor :P 2011-12-21T05:19:24 ahah ;) 2011-12-21T05:24:32 I got a question for you Janzert (yeah, again) : Were those finals prepared accurately and we're exactly where we're supposed to be. Or have those decisions been taken on the way. (The secret goal of my question is to know if you're confident about the Thursday deadline ;)) 2011-12-21T05:24:42 *** GreenTea has joined #aichallenge 2011-12-21T05:24:54 hi 2011-12-21T05:25:08 hey 2011-12-21T05:25:56 any information when cutoffs will be enabled? 2011-12-21T05:26:24 GreenTea : You have been disqualified in order to make 13130 people happier (Assuming you were supposed to end up 10th). Sorry. 2011-12-21T05:26:36 Janzert said in approx one hour 2011-12-21T05:26:49 5000 bots but not all together 2011-12-21T05:27:08 And there's the cutoff info on the server infos webpage 2011-12-21T05:27:30 i just went up a little 2011-12-21T05:27:31 :D 2011-12-21T05:27:35 *** tob__k has joined #aichallenge 2011-12-21T05:27:37 ok 2011-12-21T05:28:01 5000 bots remaining after cutoff? or 5000 bots to cutoff? 2011-12-21T05:28:05 http://tcpants.com/player/Accoun12301 let play 2011-12-21T05:28:22 *** Akranis has joined #aichallenge 2011-12-21T05:28:24 Surya: I'm don't quite understand the first portion, but as to the point of the question. there is some discussion amongst the organizers about extending the finals length 2011-12-21T05:28:34 are you by any chance discussing the cutoff rules? 2011-12-21T05:28:59 I wondered (a) when the first (second?) cutoff will take place, and at what value of mu... 2011-12-21T05:30:19 *** Jak_o_Shadows has quit IRC (Remote host closed the connection) 2011-12-21T05:30:23 *** ikaros has joined #aichallenge 2011-12-21T05:31:29 Janzert: Sorry for my poor english. I'm extremely curious on how you guys organized this challenge. And I'm wondering if you expected this amount of entries (so if everything was set up or if you have to make crucial decision all the time, in panic mode :)) 2011-12-21T05:32:00 some of the former some of the latter :) 2011-12-21T05:33:22 and anoher timeout 2011-12-21T05:33:25 wtf 2011-12-21T05:33:46 time to test it 2011-12-21T05:34:05 TL__ 5k to cutoff apparently 2011-12-21T05:34:08 this is not normal 2011-12-21T05:35:21 so the 5k bots with the lowest mu will be cutoff... an do you already know when? just interested... 2011-12-21T05:36:08 bots below rank 5000 will be cutoff 2011-12-21T05:36:12 *** katebus has joined #aichallenge 2011-12-21T05:36:26 *** Belerafon_L has left #aichallenge 2011-12-21T05:36:43 will there be more cutoff waves? 2011-12-21T05:36:59 excuse me, but I read in the forum that the cutoff will be determined by mu, while the rank is determined by the skill (mu-3*sigma) 2011-12-21T05:37:27 whoever posted that was wrong then 2011-12-21T05:37:39 okay. 2011-12-21T05:37:56 tob_k: that was just a suggestion in the forum 2011-12-21T05:38:21 (I think, the idea was to use mu so that bots with a low number of games are not at a disadvantage) 2011-12-21T05:38:56 thanks for the information! 2011-12-21T05:41:08 ok this is not normal 2011-12-21T05:41:55 *** codetiger has joined #aichallenge 2011-12-21T05:42:22 yahoooo!!! My bot is in top 100... hope it holds its position 2011-12-21T05:42:31 aichallenge: janzert epsilon * r23f54d0 / manager/cutoff_adjuster.py : Small fixes for cutoff adjuster - http://git.io/fzbDXg 2011-12-21T05:43:30 oo 2011-12-21T05:43:43 so it will be introduced in the near time 2011-12-21T05:44:11 at least there is a possibility 2011-12-21T05:44:12 *** sigh has joined #aichallenge 2011-12-21T05:44:31 and here we go 2011-12-21T05:44:41 :O 2011-12-21T05:44:47 cutoff is added now? 2011-12-21T05:44:57 http://aichallenge.org/visualizer.php?game=296698&user=757&turn=470&row=68&col=108 bah, movement bug in narrow "corridors" 2011-12-21T05:44:57 cutoff will adjust to 5000 over the next hour 2011-12-21T05:44:59 ah 13 hrz 43 mins since last game :D 2011-12-21T05:45:39 *** meduza has joined #aichallenge 2011-12-21T05:45:55 Fluxid, very defensive :) 2011-12-21T05:46:09 wow, I love the cutoff work... :) Atleast good bots get more games 2011-12-21T05:46:52 *** olexs has joined #aichallenge 2011-12-21T05:46:56 the Greedy bot is at position 2126 http://aichallenge.org/profile.php?user=12359 ;-) so the cutoff is defensive, but it is good 2011-12-21T05:48:03 if greedy is fixed to avoid timeout, it should land in within 2000 2011-12-21T05:48:19 better defensive and adjust later if needed, we will se how it will influence the contest 2011-12-21T05:48:42 I still see a lot of games with "food not gathered" within rank 3500 2011-12-21T05:48:47 *** Hexren has joined #aichallenge 2011-12-21T05:48:52 *** Hexren has left #aichallenge 2011-12-21T05:50:27 i wonder how it would be possible to provide test bots but to prevent those test bots from being submitted... 2011-12-21T05:50:50 maybe provide test bots only in binary and require source files to be submitted? 2011-12-21T05:51:24 people will still submit starter codes :D 2011-12-21T05:51:31 tob__k: those are not test bots being submitted 2011-12-21T05:51:42 this got over on 18th right? 2011-12-21T05:51:51 greedy bot? 2011-12-21T05:51:57 and greedy bot was uploaded with aggreement of organizers 2011-12-21T05:52:17 oh, okay. :-) 2011-12-21T05:52:40 well it is good that we have a few sample bots to compare to (lets say from 1 to 4) 2011-12-21T05:53:02 it gives us a rank comparison for cutoff as well 2011-12-21T05:53:32 no more bot uploads allowed now right? 2011-12-21T05:54:27 oh yeah 2011-12-21T05:54:32 sry for tht 2011-12-21T05:54:48 @codetiger, yes games around 3500 looks funny -> http://aichallenge.org/visualizer.php?game=302685&user=13614 2011-12-21T05:54:49 katebus: Run as fast as you can and don't look back. 2011-12-21T05:54:52 my impression was that 50% or more of the submissions are starter bots (but no test bots i have to admit)... maybe next year one could change the tutorial not to encourage people to submit their starter bot as the first step, before even trying any improvements. 2011-12-21T05:55:00 My bot has a very good wining % in multihill close neighbor maps... what percentage of maps/games are usually in that range 2011-12-21T05:55:05 6% cut off 2011-12-21T05:55:59 games around 3500 looks funny -> http://aichallenge.org/visualizer.php?game=302685&user=13614 2011-12-21T05:56:01 katebus: this way a bot which starts to do something lost because of timeout... 2011-12-21T05:56:10 janzert: where can we see how many cutoffs 2011-12-21T05:56:23 flowenol: server statistics 2011-12-21T05:56:30 server statistics page 2011-12-21T05:56:30 I fixed a bug to make my bot perform better in close neighbor maps, and now it works well only in that maps. Random maze has become a hell for my bot now 2011-12-21T05:56:46 Fluxid: pairint cutoff? 2011-12-21T05:56:53 i think so 2011-12-21T05:57:10 7% cut off 2011-12-21T05:57:13 :D 2011-12-21T05:57:28 oh 7317 cut-off, and it moves 2011-12-21T05:57:34 yes, pairing cutoff is it 2011-12-21T05:57:35 http://tcpants.com/player/Accoun12301 let play 2011-12-21T05:57:55 the number is how many are still up right? 2011-12-21T05:58:41 yes, how many up 2011-12-21T05:58:48 Accoun: i let you play 2011-12-21T05:58:56 starter botz ruined the real essence of final tournament :) 2011-12-21T05:59:07 and it goes down, janzert said that it will go down to 5000 2011-12-21T05:59:32 *** amstan_ has quit IRC (Ping timeout: 252 seconds) 2011-12-21T06:01:25 another suggestion for the future: maybe one could provide a simple logging feature for the starter bot, since status information can not simply be put to console. first thing I did to my bot both last year and this year was to write a simple file logger. without something like this its impossible to debug the bot and many players will quit in frustration, leaving only their starter bot to play... 2011-12-21T06:03:03 I have a friend who updated the java advanced starter pack, and he's 1792 right now oO 2011-12-21T06:03:32 *** xyxan has quit IRC (Ping timeout: 258 seconds) 2011-12-21T06:03:37 This starter pack was way too good imo 2011-12-21T06:03:59 *** denes has joined #aichallenge 2011-12-21T06:05:08 12% cutoff 2011-12-21T06:05:27 suddenly, game rate fell from 15 to 9 2011-12-21T06:05:50 suddenly i started getting more games :D 2011-12-21T06:06:21 i've had 8 games over night 2011-12-21T06:06:30 and landed around 50th 2011-12-21T06:07:11 woow just played 2 games and Next game should be within 366 minutes. 2011-12-21T06:07:20 it is at least twice as fast now 2011-12-21T06:08:01 the cut-off works 2011-12-21T06:08:35 out of 13K players cut off is 6K which means 50% of bots are marked inactive? 2011-12-21T06:08:49 Fluxid: what do you mean by 12% 2011-12-21T06:09:08 Only about 7800 actual submissions 2011-12-21T06:09:39 *** nek has joined #aichallenge 2011-12-21T06:09:44 so about 2800 are being cut off 2011-12-21T06:09:45 out of 7800, right now you have 6688 active, the rest is cut off 2011-12-21T06:09:59 codetiger: account != player, there are lots of accounts without submission 2011-12-21T06:10:22 smiley1983 : 7800 only? does that mean the rest did not submit or their bot didnt compile? 2011-12-21T06:10:25 oh ok 2011-12-21T06:10:25 as far i understand, bots with rank up to $PAIRING_CUTOFF will continue to play 2011-12-21T06:10:51 I understand it the same as fluxid 2011-12-21T06:11:56 *** alc has quit IRC (Ping timeout: 248 seconds) 2011-12-21T06:12:23 so ~16% of bots are not playing anymore 2011-12-21T06:12:23 *** xyxan has joined #aichallenge 2011-12-21T06:12:40 correct 2011-12-21T06:14:16 may be you should plan some other method for cutting off not average bots... Like in ants, food not gathering for 150 turns, that bot gets penalty 2011-12-21T06:15:40 it's a form of elimination tournament, slowly the less performant bots are eliminated 2011-12-21T06:16:45 Anyway, am happy it works now... Just that I've to wait 2011-12-21T06:17:44 haha... Rank 12 if you sort the first page in ascending number of games. Let me see how long my bot holds this postion 2011-12-21T06:18:14 hmm I wonder (just wonder) how it influences the last games of the weak bots on the border of cutoff rule, these will be paired with stronger as you cannot find weaker 2011-12-21T06:19:03 anyway for these week bots, it just the luck that counts 2011-12-21T06:19:35 Anyway, losing to a strong opponent doesn't reduce your skill much 2011-12-21T06:19:51 sure 2011-12-21T06:21:14 ane the cuott goes so fast through the list that it will not matter for these weaker bots 2011-12-21T06:21:22 relevant to that; the cutoff is soft in that a bot can get pushed below it from losing a game allowing a different one to pop back above the cutoff and take its place 2011-12-21T06:21:39 Does the matching, tries to keep number of games equal? atleast everytime before cutoff increases? 2011-12-21T06:21:43 only really happening while the cutoff is stationary of course 2011-12-21T06:22:38 codetiger: pretty much, so bots that pop back above may get a few games quickly 2011-12-21T06:23:05 you'll see quite a bit of churn around the cutoff 2011-12-21T06:23:26 how long will the cutoff be kept fixed 2011-12-21T06:23:29 janzert: Good effort :) Thanks for all this sleepless nights 2011-12-21T06:23:38 *** mcstar has joined #aichallenge 2011-12-21T06:23:43 flowenol: we'll see ;) 2011-12-21T06:23:57 15 hrz :( waiting waiting waiting :P 2011-12-21T06:24:00 kk 2011-12-21T06:24:08 I don't expect it to stay at 5000 for too long, maybe a few rounds of games 2011-12-21T06:24:11 i need moar games, othwewise my bot wont rise to its proper palace :( 2011-12-21T06:24:23 i would kill to see how my bot ends 2011-12-21T06:24:27 then probably 3000ish will be the next stop 2011-12-21T06:24:39 im still ok with that 2011-12-21T06:24:57 i will only be worried with cuts above 1000 2011-12-21T06:25:19 was there a cut already? 2011-12-21T06:25:31 it'll be a while before it gets there :) 2011-12-21T06:25:45 mcstar: in progress yes 2011-12-21T06:25:53 fantastic 2011-12-21T06:27:23 janzert: what does the recent column mean? 2011-12-21T06:27:29 recent games in what time? 2011-12-21T06:27:29 janzert: On an unrelated note, I wonder, is it actually a weakness of TrueSkill that it tries to compensate for abilities changing over time, where that's not really the case with bots? 2011-12-21T06:28:13 mcstar: Games in the last 24 hours, I think 2011-12-21T06:28:20 *** Zaphus has joined #aichallenge 2011-12-21T06:28:37 k 2011-12-21T06:28:41 matthewd: it still needs to account for the population changing and the bots performance against it 2011-12-21T06:28:50 which basically comes out to the samething 2011-12-21T06:28:53 *** xyxan has quit IRC (Quit: Page closed) 2011-12-21T06:29:00 although not in the finals 2011-12-21T06:30:02 ser_leo poor guy, he is on the last page with 14 submissions, all time out 2011-12-21T06:30:44 janzert: Ah, right 2011-12-21T06:31:04 same people tried hard on the last pages too, i hope they will understand the cut 2011-12-21T06:31:11 a->o 2011-12-21T06:34:00 lol, it can be fun browsing those pages 2011-12-21T06:34:14 I see the cutoff has begun. Good system, gives everyone a chance initially (regardless of being a starterbot or not) and now it starts locking down 2011-12-21T06:35:11 how far up will the cutoff go ? and over how long ? 2011-12-21T06:35:41 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-12-21T06:36:34 Zaphus: It's currently heading for 5000 2011-12-21T06:36:47 janzert: Is there any way to indicate to a user that they are below the cutoff line (e.g. like the deactiviation message) ? 2011-12-21T06:37:03 'deactivated' message I mean 2011-12-21T06:38:02 5000 you say? and bot 5001 right now is "Santa Clause" - is that a coincidence ?!?! 2011-12-21T06:38:29 http://aichallenge.org/rankings.php?page=51 2011-12-21T06:39:43 now cutoff is 5262, but it will reach santa.. 2011-12-21T06:40:11 who will give me a gift if santa is deactivated and in bad mood? 2011-12-21T06:40:17 *** meduza_ has joined #aichallenge 2011-12-21T06:41:48 katebus: you should have been a nice boy and help santa make a better bot 2011-12-21T06:42:10 Janzert: you can't deactivate Santa - you'll have to stop the cutoff at 5001 :-) 2011-12-21T06:42:20 just skip him 2011-12-21T06:42:25 and continue cutting off 2011-12-21T06:43:01 *** meduza has quit IRC (Ping timeout: 258 seconds) 2011-12-21T06:43:06 i wonder if theres a Baby Jesus 2011-12-21T06:43:26 oh no, Santa has dropped to 5022 :-( that would have been awesome if he were the cutoff 2011-12-21T06:43:33 *** Speedy_Consoles has joined #aichallenge 2011-12-21T06:43:40 apparently its common name among partugese people? 2011-12-21T06:44:16 cutoff 5020 2011-12-21T06:44:28 The whole name ? Santa perhaps, but SantaClaus ? 2011-12-21T06:44:33 *** meduza_ has quit IRC (Ping timeout: 258 seconds) 2011-12-21T06:44:35 aham 2011-12-21T06:44:40 and cuoff finally 5000 2011-12-21T06:44:44 poor santa 2011-12-21T06:44:56 *** TL__ has quit IRC (Ping timeout: 258 seconds) 2011-12-21T06:44:57 i thought you have some extrasensory ability katebus 2011-12-21T06:45:11 if 23 bots near 5000 lose their games, he might get back in 2011-12-21T06:45:51 mcstar: what do you mean? 2011-12-21T06:46:03 Is the scheduler still balancing game counts (doing lower numbers first) ? 2011-12-21T06:46:57 seems no 2011-12-21T06:48:25 *** codetiger has quit IRC (Remote host closed the connection) 2011-12-21T06:48:47 katebus: you "sensed" the cutoff and reported back 2011-12-21T06:49:09 *** dici has joined #aichallenge 2011-12-21T06:49:58 the server status page just reports current value 2011-12-21T06:50:24 and I have just refreshed it 3 or 4 times ;-) when someone mentioned Santa Claus is in danger 2011-12-21T06:50:49 katebus: yeah i know, when i realized your looking at that page i said what i did 2011-12-21T06:50:57 *** pairofdice has joined #aichallenge 2011-12-21T06:50:57 and who would have thought youre a guy 2011-12-21T06:51:50 I am a bot 2011-12-21T06:56:24 i havent played a game in 15 hours 2011-12-21T06:56:33 *** cseyu01 has joined #aichallenge 2011-12-21T06:56:33 aichallenge: janzert epsilon * rf9c6b94 / (website/profile.php website/sql.php): Take pairing cutoff into consideration for profile next game display - http://git.io/qkg72w 2011-12-21T06:56:39 my bot's gonna forget how to play good 2011-12-21T06:56:43 he neeeds practice 2011-12-21T06:57:47 have u tried TCPants? 2011-12-21T06:58:14 cseyu01: whats wrong with you? 2011-12-21T06:58:19 you kept timing out 2011-12-21T06:58:27 i know i know 2011-12-21T06:58:32 why? 2011-12-21T06:58:35 works fine locally 2011-12-21T06:58:39 ok 2011-12-21T06:58:47 no clue why timing out on there 2011-12-21T06:58:54 i wish i could fix it 2011-12-21T06:59:18 my new bot kills my old bot 2011-12-21T06:59:38 old==submitted ? 2011-12-21T06:59:45 but i can't even get the tutorial bot to run on TCPants 2011-12-21T06:59:56 yep old == submitted 2011-12-21T07:00:05 *** onion42 has joined #aichallenge 2011-12-21T07:00:12 575th atm 2011-12-21T07:00:16 *** meduza has joined #aichallenge 2011-12-21T07:00:19 cseyu01, have you mofied turntime value to avoid network latency timeouts? 2011-12-21T07:00:21 but new one destroys it 2011-12-21T07:00:27 tcpants was fine for me, i had timeouts on fluxid's, but my bot run ~50ms 2011-12-21T07:00:43 nope haven't tried that 2011-12-21T07:00:51 i'll try now, ty 2011-12-21T07:00:53 it can be the main reason 2011-12-21T07:01:04 cseyu01: so you are a team? 2011-12-21T07:01:12 how many: 2011-12-21T07:01:14 ? 2011-12-21T07:01:18 yep, 2 man team 2011-12-21T07:01:29 which one is talking now? 2011-12-21T07:01:32 A or B 2011-12-21T07:01:33 regisered Dec 5th lol, righ before exams 2011-12-21T07:01:53 i guess A lol 2011-12-21T07:02:14 *** Ashoka has joined #aichallenge 2011-12-21T07:02:40 I want to do a team next time, someone to bounce ideas off 2011-12-21T07:03:18 Zaphus: you have a whole corporation to consult with :D 2011-12-21T07:03:31 http://tcpants.com/player/Accoun12301 let play 2011-12-21T07:03:55 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-12-21T07:04:14 mcstar: not so much any more, the games company I worked at went under - new job has much less AI knowledge hovering around :-( 2011-12-21T07:04:49 when did it "go under"? 2011-12-21T07:05:05 when we first talked you were still employed there, werent you? 2011-12-21T07:05:14 around spring? 2011-12-21T07:05:18 elif key == 'turntime': 2011-12-21T07:05:20 self.turntime = int(tokens[1]) 2011-12-21T07:05:32 somehting here should be fixed? 2011-12-21T07:05:38 im in ants.py 2011-12-21T07:05:53 self.turntime=500 2011-12-21T07:06:03 ok, thx 2011-12-21T07:06:03 but it doesnt matter if you dont use a timer 2011-12-21T07:06:11 i do 2011-12-21T07:06:24 but I ususally execute turns in under 100 ms 2011-12-21T07:06:47 mcstar: just over a year ago, I was there when Planet wars started, but we closed during that (so I stopped competing, still managed 400th) 2011-12-21T07:06:48 timer is just incase i get stuck in a loop 2011-12-21T07:07:15 Zaphus: we're gonna beat these hipsters next time, ok? 2011-12-21T07:07:37 mcstar: got to do something about those youngsters 2011-12-21T07:08:59 you say youngsters? are you a dinosaur? 2011-12-21T07:11:26 *** staalebk has quit IRC (Read error: Connection reset by peer) 2011-12-21T07:11:31 *** staalebk has joined #aichallenge 2011-12-21T07:11:46 *** meduza has quit IRC (Ping timeout: 258 seconds) 2011-12-21T07:11:52 well he is 2x as old as me, almost 2011-12-21T07:11:59 coming up on 50 right Zaphus ? 2011-12-21T07:12:38 yeah, faster than I'd like to admit... 2011-12-21T07:13:22 ;-) 2011-12-21T07:14:00 Stay moving so you don't fossilise 2011-12-21T07:14:32 *** meduza_ has joined #aichallenge 2011-12-21T07:16:53 *** tomb___ has joined #aichallenge 2011-12-21T07:19:27 *** harakiri has joined #aichallenge 2011-12-21T07:21:28 *** praveen__ has joined #aichallenge 2011-12-21T07:22:24 *** tob__k has quit IRC (Quit: Ex-Chat) 2011-12-21T07:23:45 *** Migi32 has joined #aichallenge 2011-12-21T07:24:39 *** sh4wn has quit IRC (Ping timeout: 268 seconds) 2011-12-21T07:27:01 yay, broke my losing streak, got a win after 5 straight last-place losses 2011-12-21T07:28:09 hm, and you didnt fall below 200 before that 2011-12-21T07:28:22 :) 2011-12-21T07:28:37 and you have a couple eliminated 2011-12-21T07:28:41 i have only 1 2011-12-21T07:28:51 im depressed, im 27x 2011-12-21T07:29:05 282 2011-12-21T07:29:23 I've lost half my games, but in most of them I wasnt alone in last place, and you do better in that situation (because you drew with all those also in last place, you can actually go up in ranking) 2011-12-21T07:29:58 *** TheLinker has joined #aichallenge 2011-12-21T07:30:23 i dont i will be rank near my TrueSkill(TM) if the finals really stop today 2011-12-21T07:30:28 ah tomorrow 2011-12-21T07:30:34 thursday 2011-12-21T07:30:46 dont think* 2011-12-21T07:30:56 *** smiley1983 has quit IRC (Ping timeout: 240 seconds) 2011-12-21T07:31:06 excuse my grammar 2011-12-21T07:31:10 Zaphus: ha ha ha ha -> "Brute Force solutions are your friend" 2011-12-21T07:31:38 aichallenge: janzert epsilon * r545bd21 / (sql/2_generate_matchup.sql sql/opponent.sql): Fix min_games query in pairing - http://git.io/Nl0Ecw 2011-12-21T07:31:58 *** smiley1983 has joined #aichallenge 2011-12-21T07:32:07 How many bots in the finals atm? 2011-12-21T07:32:53 ~5000? 2011-12-21T07:33:14 http://aichallenge.org/server_stats.php <- wow do we really have that many servers? 2011-12-21T07:33:17 ~7900 but the cutoff is at 5k 2011-12-21T07:33:41 Migi32: but the figures are pretty big 2011-12-21T07:33:55 id like to see them down to 1 2011-12-21T07:33:59 or something 2011-12-21T07:34:16 there were times when i could play almost hourly 2011-12-21T07:34:28 why cant this be the case in the finals :( 2011-12-21T07:34:42 *** sdsd has joined #aichallenge 2011-12-21T07:34:50 (it is rethorical, dont answer that) 2011-12-21T07:35:07 there were only ~750 active bots near the end, then we went to 7500 active :-) 2011-12-21T07:35:12 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-12-21T07:36:46 Next Game: 1,267,973 players are ahead. 2011-12-21T07:36:48 how much has this cutoff really saved, though? I mean, starter bots (or most weak bots) don't use any CPU whatsoever, and games between starter bots probably take only a few seconds 2011-12-21T07:37:07 There was a cutoff? 2011-12-21T07:37:12 thats why we want to see smaller cutoff 2011-12-21T07:37:53 Next game should be before the end of the universe as we know it....maybe 2011-12-21T07:37:53 And it will get smaller 2011-12-21T07:37:59 migi32: I assume the cutoff will move some more, once everyone in the 5000 has had a game or three, and it will keep doing that until ~500 are left (I'm inventing the numbers) 2011-12-21T07:38:13 remember the limit on games rate now isn't time to play the game it's database load inserting it and create pairings 2011-12-21T07:38:31 o.O 2011-12-21T07:38:33 O.o 2011-12-21T07:38:36 as game length increases I can just add more workers so the same number are finishing games at a time 2011-12-21T07:38:37 janzert: That's sad :-) 2011-12-21T07:38:46 in fact that is exactly what has been happening 2011-12-21T07:38:48 quicksort not quick enough imo 2011-12-21T07:38:55 remember? janzert ? i couldnt have guessed that :) 2011-12-21T07:39:00 <-- wishes bot was a pivot 2011-12-21T07:39:22 *** olexs has quit IRC (Quit: Leaving.) 2011-12-21T07:39:28 next contest idea: organisers supply 10,000 bots, the players have to write scheduling and ranking algorithms that everyone thinks are fair :-) 2011-12-21T07:39:39 XD 2011-12-21T07:39:41 lmao 2011-12-21T07:39:42 hahha 2011-12-21T07:40:29 so the database isnt distributed? 2011-12-21T07:40:59 no, it's mysql with myisam tables :( :( :( 2011-12-21T07:41:56 I can understand it a little bit though, the matchmaking query is a pretty beastly thing :) 2011-12-21T07:42:32 Zaphus: i had the same contest idea yesterday. i think i offended people with it. 2011-12-21T07:43:18 it is a joke, but that was lost :( 2011-12-21T07:43:29 tomb___: you mean amstan didnt take it easily? 2011-12-21T07:43:40 mcstar: i mean that, yes 2011-12-21T07:43:49 tomb___ : it's not as silly as it sounds (well, ok, it is... but...) : those are not easy problems, TrueSkill has a bucketload of PhD's behind it, as do many others - it is a hard problem to solve! 2011-12-21T07:43:55 he was bombarded with complaints the whole day 2011-12-21T07:44:15 you can understand too, that sometimes only a little bit is needed to .... 2011-12-21T07:44:17 whatever 2011-12-21T07:44:23 spill the glass? 2011-12-21T07:44:29 whats the idiom? 2011-12-21T07:44:50 mcstar: yeah, he was was totally reasonable, given the circumstances. 2011-12-21T07:44:53 *** UncleVasya has joined #aichallenge 2011-12-21T07:45:22 honestly I think the problem has been with the lack of communication to the participants, the guys doing all the hard work on the back end are under the pump, and nobody was talking to the masses. and we all see it through our own little view (i.e. what's happening to your bot) 2011-12-21T07:45:55 theres a need for a Communications Officer 2011-12-21T07:45:58 hey, it says I'm "Playing in a game right now" :D 2011-12-21T07:45:58 Zaphus: i agree. i think they need someone to volunteer to do public relations 2011-12-21T07:45:59 next time they need a spokesperson (like ChrisH stepped in at the end) 2011-12-21T07:46:15 I have 17 games though, so I don't really see why I'm allowed to play already 2011-12-21T07:46:18 but I don't mind :D 2011-12-21T07:46:49 i have had zero involvement with the administration of this contest, but i would be willing to help out on communications/public relations next time 2011-12-21T07:47:04 id like to contribute code or something 2011-12-21T07:47:09 yeah, even just a volunteer with good communications skills like ChrisH did to relay things from IRC to the forums would be great 2011-12-21T07:47:13 I can draw pictures 2011-12-21T07:47:17 *** grwip has joined #aichallenge 2011-12-21T07:47:32 it's fairly easy for me to drop a word here or glance and see a question and answer it quick 2011-12-21T07:47:39 Was I supposed to say "...and my axe!"? 2011-12-21T07:47:40 yeah, I was thinking the same (about helping on comms) - it probably needs 2 or 3 people, given the 24hr timezones it runs in 2011-12-21T07:48:05 the forums though at this stage of the contest are approached with much dread and take quite an effort to write a post :/ 2011-12-21T07:48:07 i helped a lot earlier though 2011-12-21T07:48:09 if the contest has any legal issues in the U.S., I may be able to help with those 2011-12-21T07:48:13 janzert: yes, that worked well with ChrisH 2011-12-21T07:48:18 i mean i spent too much time with these things here 2011-12-21T07:48:29 how about a blog? Whenever there are issues or decisions you've made, put that on that blog. 2011-12-21T07:48:51 a blog wouldn't be much better for me than the forums really 2011-12-21T07:49:01 *** Frantic_ has joined #aichallenge 2011-12-21T07:49:20 blog / forum thread, not a lot of difference - it's having someone who can take the time to do it right that matters (not someone buried in code and crashing servers) 2011-12-21T07:49:33 since anything formal is going to take me a half hour minimum working over the wording to try and be accurate yet not picked apart to death :} 2011-12-21T07:50:29 janzert: it's the internet, it *must* get picked apart to death or the internet isn't doing it's job :-) 2011-12-21T07:50:37 janzert: ah yeah, I get that :) 2011-12-21T07:50:38 :) 2011-12-21T07:50:57 *** Frantic_ has quit IRC (Client Quit) 2011-12-21T07:51:15 *** u_ has joined #aichallenge 2011-12-21T07:51:16 Yeah, I like IRC over forums too 2011-12-21T07:51:34 yeah i can do thi pairofdice and he now has to answer me 2011-12-21T07:51:43 but if i just write something to a forum... 2011-12-21T07:51:47 problem with IRC is that if you weren't there the minute the thing was said, you miss it (more or less) 2011-12-21T07:52:12 we need a @sticky 2011-12-21T07:52:13 Yeah, but it gives you a better sense of .... concurrency 2011-12-21T07:52:18 *** u__ has joined #aichallenge 2011-12-21T07:52:36 and somone to update the sticky XD 2011-12-21T07:52:56 best solution is that IRC is where the decisions get made, then someone translates that into a forum post (sticky, perhaps no replies in some cases) 2011-12-21T07:53:35 what happened to RSS feeds? 2011-12-21T07:53:45 yeah, I agree. But I'm not sure that so many people read the forum announcements 2011-12-21T07:53:45 they were in fashion for a while 2011-12-21T07:53:49 and they disapperaed 2011-12-21T07:53:50 yes, the way ChrisH was able to relay some things really is ideal from my perspective more of that by him and or others would be great 2011-12-21T07:53:55 were they really useless? 2011-12-21T07:54:11 *** Speedy_Consoles has quit IRC (Quit: KVIrc 4.0.1 Insomnia http://www.kvirc.net/) 2011-12-21T07:54:12 I use google reader every day 2011-12-21T07:54:25 pairofdice: whats that? 2011-12-21T07:54:27 google reader's my homepage :) 2011-12-21T07:54:30 feed-reader? 2011-12-21T07:54:34 Yeah 2011-12-21T07:54:36 Once the response is out there in the forum, others point people to the answer/announcement 2011-12-21T07:54:55 Bundled with other glorious google software 2011-12-21T07:55:28 when i think about communication i think about an email-client 2011-12-21T07:55:37 janzert: would we be expecting the game time per server to be increasing as we reduce the list, and the failure rate to be decreasing ? 2011-12-21T07:55:42 and a bundled news-reader 2011-12-21T07:55:49 You can do it all in a browser 2011-12-21T07:55:51 *** u_ has quit IRC (Ping timeout: 252 seconds) 2011-12-21T07:55:52 *** u__ is now known as u_ 2011-12-21T07:55:56 Zaphus: somewhat 2011-12-21T07:56:04 pairofdice: ok, im just not used to do it in a browser 2011-12-21T07:56:32 the high game time right now is because several 10 player maps are "in vogue" with the pairing algorithm at the moment ;) 2011-12-21T07:56:47 midnight here, better head off to bed. any idea when the next round of cutoff movement is going to kick in ? 2011-12-21T07:57:03 lets make this clear 2011-12-21T07:57:05 Minutes per game per server 2011-12-21T07:57:14 *** xyxan has joined #aichallenge 2011-12-21T07:57:16 according to my interpretation, the lower the bter 2011-12-21T07:57:17 better 2011-12-21T07:57:18 I haven't made a firm decision yet 2011-12-21T07:57:28 at least a few hours 2011-12-21T07:57:38 *** Accoun has quit IRC (Ping timeout: 240 seconds) 2011-12-21T07:57:40 mcstar: yup 2011-12-21T07:57:59 next round of cut off must be quick. anybody mind cutoff set to 2000. am sure nobody out there with rank 2000+ still following their bot status 2011-12-21T07:58:01 mcstar: yes, lower will be better, but I imagine the bots will be the ones that use more of their allotted time, and live longer (than the ones being cut off) - so games will take longer 2011-12-21T07:58:03 Zaphus thinks otherwise 2011-12-21T07:58:12 k 2011-12-21T07:58:23 wow finally my Bot went up :P from 779 to 579 2011-12-21T07:59:05 when does the counter start? 2011-12-21T07:59:09 *** gustavokambara has joined #aichallenge 2011-12-21T07:59:17 when the alg. decides it needs to pair up some pplaayers 2011-12-21T07:59:24 or after it actually did the selection? 2011-12-21T07:59:28 Do you still have the stats for before the finals in the DB somewhere ? would be interesting to look to see if any bots below 2000 were in the top (say) 500 before the finals - if so it might be too soon to cut them off 2011-12-21T07:59:34 mcstar: what counter? 2011-12-21T07:59:47 *** TheLinker has quit IRC (Quit: Bye) 2011-12-21T07:59:51 the one that count game-time 2011-12-21T07:59:53 i was clueless about what was going on until i started browsing the IRC logs 2011-12-21T07:59:56 *** tomb___ has quit IRC (Quit: Page closed) 2011-12-21T08:00:00 that makes the server statistics 2011-12-21T08:00:25 janzert: is the DB overhead counted in it? or no 2011-12-21T08:00:31 the game time for the server calculation is just 30 / number of games in past half hour 2011-12-21T08:00:46 so yeah all overhead is in it 2011-12-21T08:01:07 then i expect numbers to get even lower with time 2011-12-21T08:01:13 with more aggressive cutoffs 2011-12-21T08:01:28 this went to Zaphus 2011-12-21T08:01:28 @rankings 2011-12-21T08:01:29 Ashoka: Top 10 players: xathis(85.5), meduza(83.5), FlagCapper(82.7), BenJackson(81.6), _Atsky_(81.6), teapotahedron(81.4), Speedy_Consoles(80.2), runevision(80.2), delineate(80.0), lazarant(79.7) 2011-12-21T08:01:41 @rankings 50 2011-12-21T08:01:41 Ashoka: Error: Cannot do that many players. 2011-12-21T08:01:49 :-) 2011-12-21T08:02:18 @rankings 0 2011-12-21T08:02:18 onion42: Error: Cannot do that many players. 2011-12-21T08:02:38 @ranking -1 2011-12-21T08:02:39 xyxan: No! 2011-12-21T08:02:43 :P 2011-12-21T08:02:46 lol 2011-12-21T08:02:50 @rankings 0/0 2011-12-21T08:02:50 mcstar: (rankings [how many]) -- Get the top $(how many) players. 2011-12-21T08:03:13 lol 2011-12-21T08:03:13 ~rankings 3.1415 2011-12-21T08:03:18 Just so you guys know, my friend who updated the java advanced starter pack bot is currently #1793 2011-12-21T08:03:22 @rankings 3.1415 2011-12-21T08:03:23 onion42: Top 3 players: xathis(85.5), meduza(83.5), FlagCapper(82.7) 2011-12-21T08:04:05 Surya: what does that mean? 2011-12-21T08:05:06 *** Zaphus has quit IRC (Quit: Page closed) 2011-12-21T08:05:17 hm 2011-12-21T08:05:25 http://aichallenge.org/visualizer.php?game=304532&user=4513 2011-12-21T08:05:26 mcstar: It means that some starter packs (the best ones) are still in. 2011-12-21T08:05:34 xathis left 300 ants in the hive 2011-12-21T08:05:51 lol 2011-12-21T08:06:07 Surya: i thought he is ranked 1793 with his own bot 2011-12-21T08:06:16 Yeaah, is that really a starter bot :p 2011-12-21T08:06:46 Surya: you didnt say, that the starter bot is ranked 1793, but your friend is 2011-12-21T08:07:50 mcstar : exactly, but my friend uses the java starter pack. So I guess it's just an other value to take in account 2011-12-21T08:08:09 well it sucks 2011-12-21T08:08:14 thats my opinion 2011-12-21T08:08:19 on multiple level 2011-12-21T08:08:26 *** Accoun has joined #aichallenge 2011-12-21T08:08:37 first of all, people should not have that strong starter bots 2011-12-21T08:08:51 second, it said that it still so high 2011-12-21T08:09:07 yep, I wish we could have avoided starter pack at the end. But once again, this java starter bot was way too good imo 2011-12-21T08:09:07 third, my rank will not mean much, if a starter can reach so high 2011-12-21T08:09:35 mcstar : agreed 2011-12-21T08:09:57 *** praveen__ has quit IRC (Quit: Page closed) 2011-12-21T08:10:07 Surya: your friend is raph_m? 2011-12-21T08:10:54 jiijax? 2011-12-21T08:11:26 neither one is a starter bot though 2011-12-21T08:11:39 janzert : jijaxxx 2011-12-21T08:12:09 ok, it's not a starter bot though 2011-12-21T08:14:49 janzert : ok, what I mean is he copied paster the code from the website (the 5 tutos) 2011-12-21T08:14:59 and just uploaded it. My bad 2011-12-21T08:15:06 ahh, tutorial bot 2011-12-21T08:15:20 a starter bot is much dumber :} 2011-12-21T08:15:21 yep, my bad, I thought it was already inside the Java Starter pack. 2011-12-21T08:15:23 ok :) 2011-12-21T08:15:46 not that a tutorial bot is terribly smart :P 2011-12-21T08:16:09 thats the point 2011-12-21T08:16:13 and python tutorial bots will time out continually 2011-12-21T08:16:14 it is ranked too high 2011-12-21T08:16:19 with some small modifications the tutorial boit can be pretty strong 2011-12-21T08:16:37 Keyword, modifications 2011-12-21T08:16:52 mcstar: I'm not so sure it is 2011-12-21T08:16:54 just have to get rid of "is_occupid" or whatever 2011-12-21T08:17:15 better to always use passable and weed out collisions 2011-12-21T08:17:15 *** Murashka has joined #aichallenge 2011-12-21T08:17:41 janzert: so how many people really competed in PW? 2011-12-21T08:17:51 people sometimes quote 7k 2011-12-21T08:17:54 *** UncleVasya has quit IRC (Remote host closed the connection) 2011-12-21T08:17:56 but thats bs obviiously 2011-12-21T08:18:21 so like 4.5k people submitted 2011-12-21T08:18:24 there were 7200 user accounts, 4619 submissions in the finals 2011-12-21T08:18:43 but i recon, much of those were very simple constructions 2011-12-21T08:18:45 I don't remember how many were starter bots if you wanted to weed them out as well ;) 2011-12-21T08:18:55 yes 2011-12-21T08:19:00 i wanted some approximation 2011-12-21T08:19:08 if the figures hold for this contest too 2011-12-21T08:19:14 to see if* 2011-12-21T08:19:57 http://dl.dropbox.com/u/9140666/elo-draw-chart.png 2011-12-21T08:20:09 because if a simple bot reaches 1700 from 7k bots, then id say, PW had 5-600 good contestants 2011-12-21T08:20:21 where the big green bump of draw percentage is was all the starter bots 2011-12-21T08:20:36 *** goffrie_ has joined #aichallenge 2011-12-21T08:20:36 *** goffrie has quit IRC (Disconnected by services) 2011-12-21T08:20:37 *** goffrie_ is now known as goffrie 2011-12-21T08:20:43 so roughly 2900 bots were better than starter bots 2011-12-21T08:21:16 *** ablebaker99 has joined #aichallenge 2011-12-21T08:21:30 *** gustavokambara has quit IRC (Quit: Page closed) 2011-12-21T08:21:56 *** ablebaker99 has left #aichallenge 2011-12-21T08:22:09 :P so reduce cutoff to 2500 :D 2011-12-21T08:22:24 so the gap is at ~4000 as i thought 2011-12-21T08:22:35 a1k0n: is your bot has a known weakness in 2 player maps.... ? my bot seems to be having a good time against your one at tcpants bit strange as the same bot is doing so poorly in the official server. 2011-12-21T08:22:41 Maybe the tutorial shouldn't have contained code. just pseudo code. 2011-12-21T08:23:15 Surya: python is supposed to read like pseudocode :D 2011-12-21T08:23:41 not to mention the python code given will just create a bot that times out in most games 2011-12-21T08:23:46 I like that there was code in the tutorial 2011-12-21T08:23:57 a1kon: ah part of the reasoning may be in the tutorial map, my one seems to like it so much. 2011-12-21T08:24:00 Gives a decent idea 2011-12-21T08:24:02 good to have code in the tutorial imo 2011-12-21T08:24:14 <-- leant pythong for this 2011-12-21T08:24:15 sdsd: yeah, I do think the tutorial was quite nice for people this time 2011-12-21T08:24:42 janzert :P can you give timing details for all bots :P whos is much efficient? and also in high rank?? 2011-12-21T08:24:50 i don't think i could've picked up python so fast without the code in teh tutorial 2011-12-21T08:25:05 ^same 2011-12-21T08:25:30 hm 2011-12-21T08:25:36 python is deeper than that though 2011-12-21T08:25:51 ? 2011-12-21T08:25:57 Yea, but you've got to start somewhere 2011-12-21T08:26:39 Better some implemented working code then start from scratch in a language that's new for you, imo 2011-12-21T08:26:50 xyxan: no that isn't recorded 2011-12-21T08:26:50 i dont agree sorry 2011-12-21T08:27:07 it nice when you figure out these things yourself 2011-12-21T08:27:13 of course if you have the time... 2011-12-21T08:27:15 Ok, so it's needed. My problem is that I wish I could have a proper ranking at the end. Like "you ended up" X / Y "real contesters, not starter pack or tutorial copy paster if you know what I mean. But if it helped you a lot, I guess it was a good thing :) 2011-12-21T08:27:31 hummm okk 2011-12-21T08:27:41 *** nek has quit IRC (Ping timeout: 240 seconds) 2011-12-21T08:28:10 I figured more stuff out when I started profiling the code and rewrote inefficient parts of the started bot. :) 2011-12-21T08:28:14 it took me a while to figure out the list thing 2011-12-21T08:28:16 sdsd: i might not rank well, but my bot is pretty fast at least :) 2011-12-21T08:28:24 ((...)) 2011-12-21T08:28:25 the engine doesn't measure how much time you used, just that you returned a turn within the time limit 2011-12-21T08:28:41 sometimes 3 brackets, sometimes 2 2011-12-21T08:28:49 ^(^_^)^ 2011-12-21T08:29:00 tuple unpacking 2011-12-21T08:29:05 ^this 2011-12-21T08:29:14 cseyu01: you need a tour in the world of lisp 2011-12-21T08:29:16 I love the server statistics webpage! 2011-12-21T08:29:22 Full of nice infos 2011-12-21T08:30:04 Sry4TimeOut 2011-12-21T08:30:17 i paired with this guy on tcpants 2011-12-21T08:30:20 im* 2011-12-21T08:30:34 seems he cant start a game 2011-12-21T08:30:41 I need tours in alot of worlds,I'd like to learn go 2011-12-21T08:31:11 that'd be me 2011-12-21T08:31:14 My bot is quite offensive & not that good in combat. I think the cutoff thing will make me play against good or equivalent players who will destroy me even if I'm better than them on big maps where there are noobs. 2011-12-21T08:31:22 i just closed the conn 2011-12-21T08:31:30 cseyu01: its happening again like yestreday :) 2011-12-21T08:32:05 i'm about as frusrated as a midget with a yo-yo 2011-12-21T08:32:42 the proper term is little people 2011-12-21T08:32:45 not midget 2011-12-21T08:32:48 XD 2011-12-21T08:32:49 http://tcpants.com/player/Accoun12301 let play 2011-12-21T08:33:21 janzert: top 1000 deserve atleast a full day to compete against top botz 2011-12-21T08:36:09 *** sigh has quit IRC (Remote host closed the connection) 2011-12-21T08:36:41 No they don't 2011-12-21T08:37:01 *** u_ has quit IRC (Quit: u_) 2011-12-21T08:37:09 They get bread and water 2011-12-21T08:37:48 i wish i get some pizza :O 2011-12-21T08:38:42 and fish 2011-12-21T08:38:46 dont forget the fish 2011-12-21T08:39:40 ok pizza with fish toppings :D 2011-12-21T08:40:04 why fish? 2011-12-21T08:41:30 *** valydo has quit IRC (Ping timeout: 252 seconds) 2011-12-21T08:43:08 why bread an water? 2011-12-21T08:43:40 why not pizza? :P 2011-12-21T08:44:26 did cutoff happen over night ? 2011-12-21T08:44:51 besh: the world is round 2011-12-21T08:44:55 *** FranPeruzzi has quit IRC (Ping timeout: 258 seconds) 2011-12-21T08:44:57 Just put a timebomb in starters :)) So they'll start failing at finals date :)) 2011-12-21T08:44:59 in faact, it is spinning 2011-12-21T08:45:57 map selection on tcpants sucks 2011-12-21T08:47:50 mcstar: many 2 player maps isn't it. and the tutorial.map as well. for some reason, my bot is loving them! 2011-12-21T08:48:07 i can see that 2011-12-21T08:48:13 im yet to beat you 2011-12-21T08:48:17 maybe i wont :( 2011-12-21T08:48:34 do you beat accoun? 2011-12-21T08:48:36 I have 4-2 against a1kon who is ranked #1 there. 2011-12-21T08:48:49 havn't played with Accoun yet 2011-12-21T08:50:05 mcstar: I wish this bot performs at least 1/10 the that well in the official site where it matters :-( again for some reason, it is doing very poorly there. 2011-12-21T08:50:27 50 2011-12-21T08:50:34 thats is not very bad 2011-12-21T08:50:54 id be happy to trade palaces :D 2011-12-21T08:51:17 maybe when you get more games, youll crawl up 2011-12-21T08:51:24 thats what im counting on, obviously too 2011-12-21T08:51:33 mcstar : hmm I used to be better, but then again that was some time back. I have messed something very badly in my code during last week :-( 2011-12-21T08:51:36 *** dmj111 has quit IRC (Ping timeout: 240 seconds) 2011-12-21T08:52:11 Ashoka: i changed quite much in the last 2 days, and didnt have time to properly test it 2011-12-21T08:52:24 so it can be actually worse than my previous one 2011-12-21T08:53:24 i have a switch that lets me kill still opponents 2011-12-21T08:53:34 its possible that it isnt strict enough 2011-12-21T08:53:38 and i lose ants because of that 2011-12-21T08:54:57 HEY 2011-12-21T08:55:07 ive got a notification after 17 hours 2011-12-21T08:55:18 :O 2011-12-21T08:55:18 64.86->65.97 2011-12-21T08:55:55 finally i beat some losers 2011-12-21T08:56:06 lolx 2011-12-21T08:56:18 ur rank? 2011-12-21T08:56:41 256 2011-12-21T08:56:44 :( 2011-12-21T08:57:06 Agent Smith? 2011-12-21T08:57:08 *** UncleVasya has joined #aichallenge 2011-12-21T08:57:18 no space 2011-12-21T08:57:41 uu had a match with me 2011-12-21T08:57:42 http://aichallenge.org/visualizer.php?game=282689&user=12 2011-12-21T08:57:43 :D 2011-12-21T08:57:47 *** AFKMODE is now known as BEASTMODE 2011-12-21T08:58:54 xyxan: do you have combat code? 2011-12-21T08:59:15 nope :D dat was self defence part :D 2011-12-21T08:59:15 i pretty much cleared that map 2011-12-21T08:59:21 thought so 2011-12-21T09:00:10 run time decisions make that combat :D 2011-12-21T09:00:23 if (his_crew >= my_crew): score = score - PENALTY 2011-12-21T09:00:30 ^^my combat code 2011-12-21T09:02:04 *** keith__ has joined #aichallenge 2011-12-21T09:03:25 does anybody else witnessing some strange timeouts? My bot has not timed out in the official site before (well, at least not that I remember) but already twice in the finals. this one is particularly strange. I have just 44 ants! got time checks all over the code. strange and sad. http://aichallenge.org/visualizer.php?game=304726&user=66 java bot. 2011-12-21T09:04:18 someone reported something simialr 2011-12-21T09:04:20 both times it timed out, I was doing well :-( oh well. It's just a game .............. 2011-12-21T09:04:21 ye4streday? 2011-12-21T09:04:37 there has been a lot of talk about timeouts in the finals -- from bots who believe they should not be timing out 2011-12-21T09:04:46 my bot timed out twice 2011-12-21T09:04:54 my very first game in the finals and another one today 2011-12-21T09:05:05 Ashoka: dont tell me its "just a game", you who spent 2 months with PW after it was done! 2011-12-21T09:05:16 65.93->67.05 2011-12-21T09:05:51 mcstar : congrats.. that's progress :-) 2011-12-21T09:06:08 mcstar : what's ur MU ? 2011-12-21T09:06:11 whos this? http://aichallenge.org/profile.php?user=12368 2011-12-21T09:06:38 agent_smith beat me, i came only sencond to him :D 2011-12-21T09:06:51 i'm not quite sure it's not a bug. but most probably it's a gc related issue 2011-12-21T09:06:54 so, does anyone have any feeling for whether or not the current rankings are indicative of actual skill? or are they still highly volatile? 2011-12-21T09:06:58 Ashoka: 73.69, why? 2011-12-21T09:07:21 keith__: i think they are 2011-12-21T09:07:42 *** mayanks43 has left #aichallenge 2011-12-21T09:07:54 ranking is gud 2011-12-21T09:09:12 cool thanks xyxan, mcstar 2011-12-21T09:09:55 http://tcpants.com/replay.19278 2011-12-21T09:10:02 i beat accoun, badly 2011-12-21T09:10:10 thats 1 in a 1000 2011-12-21T09:10:44 meduza_: did it timed out when you had many ants or did it happen bit strangely like in that game I have posted. ? 2011-12-21T09:10:50 damn, i'm being beat 2011-12-21T09:11:52 Fluxid: do you have snow there? 2011-12-21T09:11:57 *** Areks has quit IRC (Ping timeout: 268 seconds) 2011-12-21T09:12:18 *** dmj111 has joined #aichallenge 2011-12-21T09:13:46 http://tcpants.com/replay.19284 2011-12-21T09:13:47 omg 2011-12-21T09:13:51 im soo stupid 2011-12-21T09:14:15 i have a function that pulls back an ant near the hill, if it is the closest and tried to leave 2011-12-21T09:14:31 Ashoka: both times bot was losing and both times there were only few ants (<50) 2011-12-21T09:14:34 but i didnt give it any other criteria 2011-12-21T09:14:45 I am not fluxid but it is snowing right now here where I am 2011-12-21T09:14:52 for example: after only turn 10 or something 2011-12-21T09:15:01 i dont know where Fluxid is 2011-12-21T09:15:08 he is polsih too 2011-12-21T09:15:22 Poland, yeah, it snowed a little last night 2011-12-21T09:15:30 mcstar: too? 2011-12-21T09:15:47 meduza_: you deserve a pat on the back for being #2 even after 2 time outs :-) 2011-12-21T09:16:08 Fluxid: like katebus 2011-12-21T09:16:19 and its snowing here 2011-12-21T09:18:15 if(myants_list.size()<3){return;} 2011-12-21T09:18:23 thats the auxiliary condition 2011-12-21T09:18:39 <3 is not an emoticon 2011-12-21T09:19:17 this is some serious flaw 2011-12-21T09:19:25 mcstar: In that game with Accoun where u have won, ur bot was sparing the last hill 2011-12-21T09:19:35 Ashoka: actually i don't think this two influenced my rating much ;) both times i was losing when it happened 2011-12-21T09:19:48 *** tomb___ has joined #aichallenge 2011-12-21T09:19:54 Ashoka: i dont know how did i manage to overpower him 2011-12-21T09:19:57 meduza_:exact opposite with me unfortunately :-( 2011-12-21T09:20:44 :( 2011-12-21T09:20:54 lost 2 games :( 2011-12-21T09:21:06 *** carpcrap has quit IRC (Quit: Page closed) 2011-12-21T09:21:43 *** keith_ has quit IRC (Ping timeout: 258 seconds) 2011-12-21T09:22:06 mcstar: anything can happen in multihill maps. you have a decent bot so it is bound to win some. :-) 2011-12-21T09:22:29 thx for cheering me up 2011-12-21T09:23:20 Ugh, my bot was at #456 2011-12-21T09:23:55 Bad bot, bad! 2011-12-21T09:24:17 pairofdice: it curses its god for not giving him enough brains 2011-12-21T09:24:49 595 => 635 :( 2011-12-21T09:24:56 Next challenge I'll make it a smarter bot 2011-12-21T09:25:01 wow I'm really getting a ton of games. Awesome. 2011-12-21T09:25:33 did Totaalnet donate physical servers? Or just servers that you could use for the duration of the finals? 2011-12-21T09:25:41 *** travelan has joined #aichallenge 2011-12-21T09:25:59 Migi32: but you are falling arent you? 2011-12-21T09:26:00 They donated server time 2011-12-21T09:26:00 hi 2011-12-21T09:26:12 i m getting games with 8 9 10 botz :( 2011-12-21T09:26:28 travelan: hi 2011-12-21T09:26:30 And totaalnet servers aren't used in the finals afaik 2011-12-21T09:27:14 mcstar: I knew that my bot wasn't a top 20 bot before the finals even started. I was surprised that they were up there for a while anyway :) 2011-12-21T09:27:32 i imagined you there 2011-12-21T09:28:27 nope. If I had 2 more days, maybe. But I started writing the second version of my bot too late. 2011-12-21T09:29:42 I just hope to finish in the top 80. That means I'm the 1%! :D 2011-12-21T09:30:19 my iq puts me in the upper 2% so i'd be satisfied with that too 2011-12-21T09:30:33 i m 8% lol 2011-12-21T09:31:20 Migi32: that could be a consoling view to many. tx :-) 2011-12-21T09:31:48 mcstar : http://tcpants.com/replay.19273 2011-12-21T09:31:59 Ashoka: No timeout but in 1 or 2 games I've noticed that my bot have time only for battle and no time for strategy part. 2011-12-21T09:32:22 But I turn off battle resolution when 200ms left so it must have time. 2011-12-21T09:32:44 nice 2011-12-21T09:32:48 *has 2011-12-21T09:32:56 *** avdg has joined #aichallenge 2011-12-21T09:33:21 Ashoka: yup. When I first saw all the really good submissions coming in the few days before the deadline, I was a bit sad, cause I'd been like top 10 for ages. But then I realized just making it to the first page is pretty impressive. 2011-12-21T09:33:41 Ashoka: you are good at supply, probably that part i f'd up in the end 2011-12-21T09:33:51 UncleVasya: Tx. I'm still worndering how my bot timed out at 44 ants when even a battle was not at hand. 2011-12-21T09:33:51 i run out of support ants 2011-12-21T09:35:55 what should be the next PairingCutOff? 2011-12-21T09:36:13 UncleVasya: did we plan to play online Go once? :-) 2011-12-21T09:36:40 Ashoka: I believe I've read on forums something like 'if servers are busy with other stuff there is a probability that your bot will not get its time'. 2011-12-21T09:37:11 Ashoka: I had a thought to ask same question :) 2011-12-21T09:37:39 I didnt handled timeouts in my code :P because my code even can handle 2000 ants at a time :D 2011-12-21T09:38:12 You were eashoka that time, right? Not a big change but what was your motivation for it? 2011-12-21T09:38:42 2 weeks before the finals I'm expecting to get into the top50, now I would be very happy if I could get into the top100. 2011-12-21T09:39:07 http://tcpants.com/player/Accoun12301 let play 2011-12-21T09:39:29 UncleVasya: it's shorter and is my first name :-) 2011-12-21T09:40:58 Ashoka: still have 20mb per day, I did try to look at the janzerts go site but could not launch their client. 2011-12-21T09:41:06 UncleVasya: about time out, yea I thought so too. after all, it will be really hard to avoid such an occurring once in a while. It's just my bad luck. 2011-12-21T09:41:35 UncleVasya: sorry I got totally busy after that day and totally forgot all about it. 2011-12-21T09:42:00 *** hotzenplotz has joined #aichallenge 2011-12-21T09:42:24 Am I the only one who can't load the site? 2011-12-21T09:43:34 *** Anilm3 has joined #aichallenge 2011-12-21T09:43:48 what is the most active TCP server that will stay online for some time? 2011-12-21T09:44:26 the king got his crown back 2011-12-21T09:44:35 Accoun: our record stands at 1-1 over there. and I have 5-5 with a1k0n. getting interesting. 2011-12-21T09:45:28 UncleVasya: seems like it. I can load. 2011-12-21T09:45:32 *** UncleVasya has quit IRC (Read error: Connection reset by peer) 2011-12-21T09:46:15 *** besh has quit IRC (Ping timeout: 258 seconds) 2011-12-21T09:47:34 mcstar: it is pairing u and me together all the time! 3 times in a row now 2011-12-21T09:48:05 im watching oscar acceptance speeches 2011-12-21T09:48:45 anyone? 2011-12-21T09:48:48 Someone picking a babyname? 2011-12-21T09:49:09 travelan: youre playing on it 2011-12-21T09:49:26 *** UncleVasya has joined #aichallenge 2011-12-21T09:49:39 *** xyxan has quit IRC (Quit: Page closed) 2011-12-21T09:50:08 thanks :) just discovered it 2011-12-21T09:50:58 travelan: you can download a lot of source by now, compile them yourself or whatever and you can play against them locally 2011-12-21T09:51:10 travelan: tcpants.com is one, there may be others. too 2011-12-21T09:51:43 I really want to get into shape for the next competition :) 2011-12-21T09:52:52 *** harakiri has quit IRC (Quit: Leaving) 2011-12-21T09:53:44 *** zyberkiddy has joined #aichallenge 2011-12-21T09:53:59 hmm, I notice that xathis bot jammed in this game: http://aichallenge.org/visualizer.php?game=302983&user=4513 2011-12-21T09:54:57 oh well 2011-12-21T09:58:23 wow, he came back from 3-15 2011-12-21T10:00:17 *** Murashka_ has joined #aichallenge 2011-12-21T10:00:21 *** Ashoka has quit IRC (Quit: Page closed) 2011-12-21T10:02:20 Jammed? 2011-12-21T10:02:44 *** Murashka has quit IRC (Ping timeout: 258 seconds) 2011-12-21T10:03:58 *** zyberkiddy has left #aichallenge 2011-12-21T10:05:09 *** Antimony has joined #aichallenge 2011-12-21T10:05:41 aww man, how does my bot keep losing? 2011-12-21T10:05:49 It's lost the last 3 games in a row 2011-12-21T10:06:03 thats kind of lucky 2011-12-21T10:06:57 *** Murashka_ has quit IRC (Ping timeout: 258 seconds) 2011-12-21T10:07:16 Antimony: even a preprocessor macro beat you 2011-12-21T10:07:57 *** Manwe56 has quit IRC (Quit: Page closed) 2011-12-21T10:07:59 *** Murashka has joined #aichallenge 2011-12-21T10:08:19 ? 2011-12-21T10:08:22 (also dont get me wrong, i know the feeling) 2011-12-21T10:08:26 *** Extrarius has quit IRC (Read error: Connection reset by peer) 2011-12-21T10:08:46 *** Extrarius has joined #aichallenge 2011-12-21T10:08:56 does anyone know how to get to the Server Statistics website? I can't find it :) 2011-12-21T10:09:35 http://aichallenge.org/server_stats.php 2011-12-21T10:09:58 Thanks! 2011-12-21T10:11:01 The link is at the bottom of any page 2011-12-21T10:11:30 Antimony: there are a lot of unknown names in the top x 2011-12-21T10:12:21 *** Kurnevsky has joined #aichallenge 2011-12-21T10:13:02 wow, 36 servers 2011-12-21T10:14:09 *** praveen__ has joined #aichallenge 2011-12-21T10:18:12 *** hotzenplotz has quit IRC (Quit: Page closed) 2011-12-21T10:19:43 *** Storm__ has joined #aichallenge 2011-12-21T10:20:06 mcstar: Currently working on documenting my code well. Hoping it will be useful for someone. 2011-12-21T10:20:45 *** sdsd has quit IRC (Ping timeout: 258 seconds) 2011-12-21T10:20:53 I was wondering, what does Google actually do in this challenge? Do they give money? A lot? (yeah I'm curious) 2011-12-21T10:22:24 the university recruiting people gave some money. not a lot but enough to run some EC2 instances for a while 2011-12-21T10:22:40 *** flowenol has quit IRC (Ping timeout: 258 seconds) 2011-12-21T10:23:06 jeff probably could have asked for more. 2011-12-21T10:23:20 *** denes has quit IRC (Read error: No route to host) 2011-12-21T10:24:50 UncleVasya: commenting the code or writing a blog post? 2011-12-21T10:25:34 http://tcpants.com/player/Accoun12301 let play 2011-12-21T10:25:42 *** denes has joined #aichallenge 2011-12-21T10:26:37 mcstar: 1st. I have no blog, also there are no interesting ideas in my bot to write post about. 2011-12-21T10:26:49 haha 2011-12-21T10:27:04 nobody has blogs until they do 2011-12-21T10:27:08 until they get one 2011-12-21T10:27:12 then they have 2011-12-21T10:27:37 there must be something interesting there 2011-12-21T10:28:35 UncleVasya: you should nevertheless write an outline of what your bot does on the higher level 2011-12-21T10:28:41 besides documenting the code 2011-12-21T10:28:46 yeah, i recommend installing jekyll and getting familiar with it. it's somewhere between blog-generator and raw static html here-is-some-code kind of stuff 2011-12-21T10:29:00 I'm trying to write something like game report...like it was in VGA Planets long time ago :) 2011-12-21T10:29:03 (if you have a website already) 2011-12-21T10:29:23 mcstar: I wrote some info in the beginning of the file. 2011-12-21T10:30:34 i think i have two blogs, but i dont remember their names 2011-12-21T10:31:01 As about work-flow (battle resolution --> food & hills --> ...), it is very clear from the debug strings in the turn function. 2011-12-21T10:31:07 if i end up in the top100 ill write something as well 2011-12-21T10:32:29 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-12-21T10:33:42 UncleVasya: you are really doing great 2011-12-21T10:33:54 i still cant believe you didnt use ocaml before XD 2011-12-21T10:34:10 *** katebus has quit IRC (Ping timeout: 258 seconds) 2011-12-21T10:34:12 ok, i believe, but its unbelieveable 2011-12-21T10:34:21 do you think it helped you a lot? 2011-12-21T10:34:27 or are you just smart? 2011-12-21T10:34:32 and no matter the language? 2011-12-21T10:34:33 It was very painful in June to do basic things. 2011-12-21T10:34:47 no, I just had very much time :-) 2011-12-21T10:35:47 UncleVasya: but still thinking functionally doesnt happen overnight 2011-12-21T10:36:02 and as i understand ocaml was your first functinal language 2011-12-21T10:36:26 As about any language: I know every of them (c++, Delphi, OCaml) only on the level 'conditions, loops, functions, arrays or lists' 2011-12-21T10:36:44 did you write code in ocaml the way you would in c++? 2011-12-21T10:36:55 I think it is all that one need to write entries to AIC :) 2011-12-21T10:38:00 *** Murashka has quit IRC (Ping timeout: 258 seconds) 2011-12-21T10:38:20 im gonna put your bot up for tcpants 2011-12-21T10:38:23 mcstar: I tried to follow Smiley-s style: was copying his functions, editing them. 2011-12-21T10:39:05 Bubble sort stoled from the internet :) 2011-12-21T10:39:39 *was stolen 2011-12-21T10:40:39 why did you need bubble sort? 2011-12-21T10:40:49 wasnt that part of some standard library?? 2011-12-21T10:40:54 Sometimes I didn't know how to make a thing without variables and was inserting mutable variables in my code :( 2011-12-21T10:41:04 dont tell me ocaml doesnt have a fast sorting alg. 2011-12-21T10:41:24 now I know there is a standart sorting functions :) 2011-12-21T10:41:29 *there are 2011-12-21T10:41:59 The program issued a command but the command length is incorrect. 2011-12-21T10:42:01 nice. 2011-12-21T10:42:04 I mean not after you said this, but in last few months 2011-12-21T10:42:22 the _program_ issued a command (nope, I think that was me) 2011-12-21T10:43:07 I was lurking for sort function implementation before I decided to look at the standart libraries. 2011-12-21T10:43:08 http://tcpants.com/replay.19368 2011-12-21T10:43:11 i beat accoun again 2011-12-21T10:43:27 mcstar: are you doing well on the official site? 2011-12-21T10:43:34 no 2011-12-21T10:43:36 I keep getting paired with ChrisH, memetix, etc. 2011-12-21T10:43:38 lol 2011-12-21T10:43:39 i need moar games 2011-12-21T10:43:52 yep, youre declining 2011-12-21T10:43:53 cool 2011-12-21T10:44:13 mcstar: yeah, mu is dropping *very* slowly 2011-12-21T10:44:34 I'm ending up in the middle (not first, but not last) 2011-12-21T10:44:54 do you think you gonna climb back? 2011-12-21T10:45:00 no 2011-12-21T10:45:04 im quite sure that more games will help me 2011-12-21T10:45:06 I don't think I'll drop more though 2011-12-21T10:45:21 well, I might - but I think I'll finish around 30th to 50th 2011-12-21T10:45:44 last minute submitter changed the picture quite much :) 2011-12-21T10:45:45 maybe in the 20s if I get good games/starting positions 2011-12-21T10:45:53 except leaving xathis the winner 2011-12-21T10:46:48 mcstar: http://aichallenge.org/visualizer.php?game=296700&user=2255 this was the last game where my bot didn't suck 2011-12-21T10:46:53 mu has declined since then 2011-12-21T10:47:03 anyway that was when I started losing 2011-12-21T10:47:29 UncleVasya: you just timed out on tcpants 2011-12-21T10:47:30 when all the top bots started passing me, so I got put in games with all of them a bunch of times xD 2011-12-21T10:47:44 mcstar: I can upload my c++ entry for pw. This was not the first experience in c++, but the first serious one (more than ~70 lines of code). 2011-12-21T10:48:19 the psychoterapist 2011-12-21T10:48:25 Before that I was doing only tasks for my technical school in c++: 2011-12-21T10:48:30 whoa, 36 servers 2011-12-21T10:48:41 UncleVasya: your code for pw was 70lines? 2011-12-21T10:48:54 Find max or ming entry in array, calculate average of the numbers in array and so. 2011-12-21T10:49:08 Small is beautifull 2011-12-21T10:49:16 mcstar: I mean anything before that was no more than 70 lines. 2011-12-21T10:49:28 ok 2011-12-21T10:49:31 +/ % # 2011-12-21T10:49:38 small isn't beautiful imo 2011-12-21T10:49:39 J 2011-12-21T10:49:43 lets avg 2011-12-21T10:49:56 thestinger: i finally understand that 2011-12-21T10:50:26 finally won a game after a string of last place loses 2011-12-21T10:50:59 *** codetiger has joined #aichallenge 2011-12-21T10:52:30 * avdg is surprised after watching the server page 2011-12-21T10:52:34 *** pguillory has joined #aichallenge 2011-12-21T10:54:16 hey pguillory 2011-12-21T10:54:28 your go bot is missing on github 2011-12-21T10:55:42 UncleVasya: your bot i run always times out 2011-12-21T10:56:30 doh ok 2011-12-21T10:56:39 mcstar: only via tcp (okay for local testing)? 2011-12-21T10:56:44 tcp 2011-12-21T10:56:48 *** JorgeB has joined #aichallenge 2011-12-21T10:57:05 Fluxid: lol, next time you should really use ctypes 2011-12-21T10:57:23 *** smjms has joined #aichallenge 2011-12-21T10:57:32 a1k0n: oh you have to log in on github or it shows a 404, maybe that's it? 2011-12-21T10:57:41 i'm logged in 2011-12-21T10:57:54 meh, I wonder if many people are downloading my bot just because it was listed as first 2011-12-21T10:57:56 i can see your c one 2011-12-21T10:58:13 did you maybe create it without pushing to it? 2011-12-21T10:58:18 pguillory: is that half of a G in the background? 2011-12-21T10:58:33 * avdg thinks he should add a readme 2011-12-21T10:59:09 a1k0n: oh the repo was still private, fixed it 2011-12-21T10:59:29 a1k0n: pguillory: I see the same effect 2011-12-21T10:59:31 anyway, how do you pronounce your name? like gil-roy? 2011-12-21T10:59:44 mcstar: half a g? 2011-12-21T10:59:49 in your avatar 2011-12-21T10:59:50 pguillory: ahha, thank you 2011-12-21T11:00:02 gill-or-ee 2011-12-21T11:00:18 thx, it was annoying 2011-12-21T11:00:30 when i see something i try to pronounce it in my mind 2011-12-21T11:00:33 heh np 2011-12-21T11:01:36 whoever was talking about bubblesort before. i used a bubble sort in mine too. i wrote it in c++, which has the stl sorting algorithms, but i dont use them very often and always forget the syntax 2011-12-21T11:01:49 now i'm tempted to port my 2d infinite world engine to appengine with go (here btw: http://a1k0n.net/code/jscroll/b/ ) 2011-12-21T11:02:05 tomb___: heh, I find insertion sort easier to code :P 2011-12-21T11:02:25 sometimes it is easier to just code it myself than figure out how the built-in feature works 2011-12-21T11:02:29 for bubble sort my brain has to turn on a bit 2011-12-21T11:02:38 sort(container.begin(), container.end()) :P 2011-12-21T11:02:44 :D 2011-12-21T11:02:46 ah, for some reason bubble sort is ingrained in my mind 2011-12-21T11:02:51 good old STL 2011-12-21T11:03:00 i want to write merge sort 2011-12-21T11:03:06 i haven't coded anything in two years, but bubble sort is what i can remember 2011-12-21T11:03:16 a1k0n: that looks great 2011-12-21T11:03:19 two for loops 2011-12-21T11:04:00 mcstar: a simple merge sort that uses insertion sort for runs under 64 is faster than std::sort for me 2011-12-21T11:04:06 I wrote up my symmetry detection and put up a demo for anyone interested... 2011-12-21T11:04:12 thanks. i used to have a multiplayer version with enemies and stuff 2011-12-21T11:04:12 https://github.com/jcdny/bugnuts/wiki/Symmetry 2011-12-21T11:04:28 by the way is GO language comparable to c in performance, and not like python? 2011-12-21T11:04:37 codetiger: yep 2011-12-21T11:04:41 mcstar: and running it with 2 threads makes it ~20% faster 2011-12-21T11:04:54 thestinger: i was very much surprised when it turned out the stl permutation is indeed very fast 2011-12-21T11:05:06 it beat my own alg 2011-12-21T11:05:12 *** dvladim has joined #aichallenge 2011-12-21T11:05:18 am working on a puzzle solver program which is already in python. am porting it to c to get performance. I would love to get my hands on GO 2011-12-21T11:05:32 *** Anilm3 has quit IRC (Quit: Lost terminal) 2011-12-21T11:05:34 Go is ~3x slower than C with 6g/6l 2011-12-21T11:05:41 8g/8l are a bit worse 2011-12-21T11:05:56 gccgo is faster but then you can't really use goroutines because it uses a pthread for each of them 2011-12-21T11:06:09 is l the linker or what? 2011-12-21T11:06:13 yeah 2011-12-21T11:06:24 you can cross-compile really easily 2011-12-21T11:06:43 it essentially just converts Go -> assembly (barely any optimization) 2011-12-21T11:06:46 oh ok. I've a 10X10 map and my job is to find the unique path to remove all tiles in the map. I do a lot of recursion. With Python it took 10 hrs to solve a map. :( 2011-12-21T11:07:02 *** |UncleVasya| <|UncleVasya|!kvirc@46-133-224-114.dialup.umc.net.ua> has joined #aichallenge 2011-12-21T11:07:22 <|UncleVasya|> mcstar: http://ompldr.org/vYnVscg/cpp_before_PW.zip 2011-12-21T11:07:55 Am working on puzzle generating code for this game. http://itunes.apple.com/in/app/crazy-cubes/id395501878?mt=8 2011-12-21T11:08:12 a1k0n: the standard compilers (8g, 8l, etc.) are based on the plan9 C compiler, and the goal is basically fast compilation time 2011-12-21T11:08:30 so performance is pretty disappointing for now because there isn't a mature llvm or gcc frontend 2011-12-21T11:08:52 thestinger: there are definitely some things where go is slower than C but I think its more like 50% slower not 3x. 2011-12-21T11:08:56 ah. 2011-12-21T11:09:02 ah the famous plan9 2011-12-21T11:09:12 i love the bunny though 2011-12-21T11:09:39 jcdny: have you benchmarked? it's pretty slow if you're comparing to gcc/clang with -O2/-O3 2011-12-21T11:09:43 *** UncleVasya has quit IRC (Ping timeout: 252 seconds) 2011-12-21T11:09:47 *** Speedy_Consoles has joined #aichallenge 2011-12-21T11:09:47 may be I should port to c this time. I'll keep GO for my next project. This time I might have to bring down time as much as I can 2011-12-21T11:10:01 Yes I have. 2011-12-21T11:10:02 <|UncleVasya|> mcstar: http://ompldr.org/vYnVsdw/cpp_after_PW.zip 2011-12-21T11:10:28 |UncleVasya|: what is the first archive? 2011-12-21T11:10:29 The bounds checking is ~ 10-15% for things where thats relevant. 2011-12-21T11:10:33 I can't get goroutines to scale very well over multiple cores 2011-12-21T11:10:39 there are a bunch of weird named files in it 2011-12-21T11:11:13 jcdny: yeah, but the compilers don't inline at all 2011-12-21T11:11:15 then theres gc stuff and code generation which is admittedly weak relative to a heavily optimized compiler. 2011-12-21T11:11:25 inlining is coming soon. 2011-12-21T11:11:29 and interfaces are all virtual method calls 2011-12-21T11:11:38 they are doing the work now to support it 2011-12-21T11:11:58 <|UncleVasya|> mcstar: first archive - almost all cpp I've done before PW 2011-12-21T11:12:07 *** travelan has quit IRC (Ping timeout: 258 seconds) 2011-12-21T11:12:08 oh 2011-12-21T11:12:13 pretty basic stuff 2011-12-21T11:12:38 jcdny: oh, it's gotten a lot faster with the last release 2011-12-21T11:12:42 y = sqrt(abs(a*pow(x1,2) + b*x1 + c)); 2011-12-21T11:12:52 |UncleVasya|: abs is integer absolute value 2011-12-21T11:13:02 but all the vars are floats 2011-12-21T11:13:50 comparable to Java now http://shootout.alioth.debian.org/u64/which-programming-languages-are-fastest.php 2011-12-21T11:13:51 If you look at alioth at the median speeds you get a good idea where Go falls out... http://shootout.alioth.debian.org/u64/which-programming-languages-are-fastest.php?go=on&gcc=on&calc=chart 2011-12-21T11:13:59 and imo it's a much nicer language than Java 2011-12-21T11:15:29 <|UncleVasya|> mcstar: http://ompldr.org/vYnVtMQ/Botcpp.zip 2011-12-21T11:15:31 #pragma comment(linker, "/STACK:16777216") // We need memory. More, mo-o-o-re! :-) 2011-12-21T11:16:21 *** Ashoka has joined #aichallenge 2011-12-21T11:16:24 <|UncleVasya|> MyBotStable.cpp - the very first cpp-version (I've ported my bot from Pascal because Pascal wasn't supported by contest that time). 2011-12-21T11:16:37 *** conor_f has joined #aichallenge 2011-12-21T11:16:40 http://forums.aichallenge.org/viewtopic.php?f=27&t=2205&sid=b3ac1afa821212211130b11c3749bae9 <- spam? 2011-12-21T11:16:45 jcdny: some of those benmcjarks are a bit unfair, since most languages are using gmp for pidigits and a non-native regex library for the regex one 2011-12-21T11:16:49 benchmarks* 2011-12-21T11:17:25 <|UncleVasya|> mcstar: I do not know what this string means, the only thing I know that it lets me make big arrays :) 2011-12-21T11:17:33 yeah the Go regex-dna # is terrible since they use a native one with different performance characteristics 2011-12-21T11:17:51 *** celes has joined #aichallenge 2011-12-21T11:17:53 <|UncleVasya|> Again, found it in Google (or someone's code) 2011-12-21T11:17:54 they also have a rewrite of the whole regex library in the devel branch 2011-12-21T11:18:09 Russ Cox (who made re2 https://code.google.com/p/re2/wiki/Install?tm=4) is one of the people working on Go 2011-12-21T11:18:21 oops, weird link https://code.google.com/p/re2/ 2011-12-21T11:18:32 <|UncleVasya|> mcstar: Fro fater-PW era I can give a link to the problems description. 2011-12-21T11:18:36 <|UncleVasya|> *For 2011-12-21T11:18:41 <|UncleVasya|> *after 2011-12-21T11:18:53 *** smjms has quit IRC (Quit: Konversation terminated!) 2011-12-21T11:18:54 what problems? 2011-12-21T11:19:05 anyway, i cant read code now 2011-12-21T11:19:12 ive got to clip my fingernails 2011-12-21T11:19:24 and tomorrow im travelling 2011-12-21T11:19:46 this is something they are concerned about with the regexp parser http://swtch.com/~rsc/regexp/regexp1.html 2011-12-21T11:19:52 i've found that a hindrance to writing code before but not to reading it 2011-12-21T11:20:08 <|UncleVasya|> mcstar: traveling where? 2011-12-21T11:20:19 <|UncleVasya|> To the study place (or home)? 2011-12-21T11:20:48 |UncleVasya|: to my grand-parents for christmas 2011-12-21T11:20:57 ive not been there since years 2011-12-21T11:20:59 jcdny: btw, have you looked at Rust? 2011-12-21T11:21:03 http://www.youtube.com/watch?v=zqi0DwNLJdM 2011-12-21T11:21:04 <|UncleVasya|> nice kid :) 2011-12-21T11:21:30 all users at 13 or > games 2011-12-21T11:21:40 only in passing. 2011-12-21T11:22:11 haha, seen that blond beffore 2011-12-21T11:22:29 <|UncleVasya|> mcstar: http://acm.timus.ru/author.aspx?id=110771 2011-12-21T11:22:31 jcdny: it's basically a "better C++" (where Go is a "better C") 2011-12-21T11:22:40 but, i thought Europe was a country 2011-12-21T11:22:51 so they have generic functions, algebraic data types, pattern matching, etc. 2011-12-21T11:22:56 <|UncleVasya|> ok, I'm going to comment code :) 2011-12-21T11:22:57 I am definitely in the Go camp. 2011-12-21T11:23:00 *** pguillory has quit IRC (Quit: Page closed) 2011-12-21T11:23:00 sort of like haskell + C++ 2011-12-21T11:23:11 jcdny: maybe :P 2011-12-21T11:23:21 http://www.rust-lang.org/ 2011-12-21T11:23:31 don't you miss generic functions in Go? 2011-12-21T11:23:39 I mean you can't write a generic Set data type in the language itself... 2011-12-21T11:23:47 or a tree or whatever 2011-12-21T11:23:59 their philosophy is much closer to mine. And I hate C++ with the fury of a 1000 suns. 2011-12-21T11:24:11 well then don't think of Rust as a better C++ 2011-12-21T11:24:18 but it has more ambition than just replacing C 2011-12-21T11:24:55 So to me a better C++ would include some sort of time machine to expression to go back and strangle Stroustrup in his crib 2011-12-21T11:25:09 new predictions http://pastebin.com/Tq1ekduy 2011-12-21T11:25:13 well, Rust is very similar to Go 2011-12-21T11:25:24 same kind of concurrency as Go/Erlang 2011-12-21T11:25:42 utf8 strings, local type inference, memory safe 2011-12-21T11:25:57 *** cofesor has joined #aichallenge 2011-12-21T11:26:14 but they have generic functions and data types, more functional stuff (like pattern matching) 2011-12-21T11:26:25 and immutable data by default is a great choice to make for a modern language 2011-12-21T11:26:53 winning bots http://pastebin.com/Rvdp7z8u 2011-12-21T11:27:31 data Tree a = Empty | Node a (Tree a) (Tree a) 2011-12-21T11:27:37 being able to define a type like that is awesome 2011-12-21T11:27:48 if you like types... 2011-12-21T11:28:01 http://tcpants.com/player/Accoun12301 let play 2011-12-21T11:28:13 I'm not sure what syntax Rust uses for that, but they probably just threw in a bunch of {} and ; 2011-12-21T11:28:27 *** taeke has joined #aichallenge 2011-12-21T11:29:12 hello everybody 2011-12-21T11:29:20 wow, 37 workers 2011-12-21T11:29:22 *** celes has quit IRC (Ping timeout: 258 seconds) 2011-12-21T11:29:35 McLeopold: can the matchmaking keep up with that? 2011-12-21T11:29:37 oh, and we have a pairing cutoff 2011-12-21T11:29:46 the limit is the games per minute 2011-12-21T11:29:54 we can handle about 10-11 2011-12-21T11:29:55 *** smiley1983 has quit IRC (Ping timeout: 255 seconds) 2011-12-21T11:30:00 thestinger: that look like haskell typeclass defs 2011-12-21T11:30:02 right now it is at 6.8 2011-12-21T11:30:10 mcstar: it is :P 2011-12-21T11:30:13 I dunno the Rust syntax 2011-12-21T11:30:21 i just read about it yesterday night :) 2011-12-21T11:30:38 *** smiley1983 has joined #aichallenge 2011-12-21T11:30:53 there's Data.Tree, but it shows the power of the language 2011-12-21T11:31:10 data Bool = False | True, data Maybe a = Nothing | Just a, etc. 2011-12-21T11:31:31 I would have used haskell for this, but the ghc version is so old :P 2011-12-21T11:31:57 McLeopold: Could you make your prediction thing until the 250th rank plz? :) 2011-12-21T11:32:05 k 2011-12-21T11:32:12 this bot has got 88 games http://aichallenge.org/profile.php?user=194 2011-12-21T11:32:24 Surya: :D 2011-12-21T11:32:25 thanks 2011-12-21T11:32:27 (Wanna know if you predicate i'm going to end in front of my colleague :p) 2011-12-21T11:32:50 http://pastebin.com/BYVX8Bxx 2011-12-21T11:33:06 McLeopold : Thx! 2011-12-21T11:33:22 what's the prediction based on? 2011-12-21T11:33:30 mu 2011-12-21T11:33:36 praveen__: looks like a headbanger :p 2011-12-21T11:33:56 that list is really broken 2011-12-21T11:33:57 mu + signa * 3 2011-12-21T11:33:59 im not on it 2011-12-21T11:34:02 *** Garf has joined #aichallenge 2011-12-21T11:34:52 meh, I hope that I win at least one game, I'm loosing atm 2011-12-21T11:35:06 Top 100 in the next challenge, damnit! 2011-12-21T11:35:21 (Or top 1%) 2011-12-21T11:35:29 * avdg hopes he can more concentrate on his implementation next time 2011-12-21T11:35:41 mine is still waiting for 16th. for some reason, it seems, mine always has the least number of games I can ever see on any list. Country-wise, language-wise, etc 2011-12-21T11:35:51 pairofdice: you just need to post links everywhere on reddit and we'll all get in the top 0.5% :) 2011-12-21T11:35:59 bah, i fell below 80 2011-12-21T11:36:09 it turns out i wasn't THAT great 2011-12-21T11:36:11 duh 2011-12-21T11:36:22 McLeopold, do you plan to equalize games count a bit today like yesterday? 2011-12-21T11:36:36 there aren't enough games yet to even call anyone's rank but xathis's 2011-12-21T11:37:12 yep 2011-12-21T11:37:14 i totally agree 2011-12-21T11:37:22 evern more when i see a1k0n 's bot that low 2011-12-21T11:37:27 on the predicitons i mean 2011-12-21T11:37:40 * avdg wonders how many servers will be added to the farm 2011-12-21T11:37:43 we are at 37 2011-12-21T11:38:36 there are too big changes among these predictions 2011-12-21T11:38:45 i sorry to say, but they dont mean jacks. 2011-12-21T11:39:00 mcstar: they seem pretty much the same each time to me, but I'm lower and lower :P 2011-12-21T11:39:28 i dont see convergence 2011-12-21T11:39:29 bearoff_w: we changed the seed selection permenantly 2011-12-21T11:39:40 *** smiley1983 has quit IRC (Quit: leaving) 2011-12-21T11:39:47 *** dvladim has quit IRC (Ping timeout: 252 seconds) 2011-12-21T11:40:13 McLeopold, starnge, I can't feel this ( I see two peaks of games count still 2011-12-21T11:40:44 hmm, am I a seed atm? 2011-12-21T11:42:43 game counts http://pastebin.com/LDBatDm4 2011-12-21T11:42:51 *** moxian has joined #aichallenge 2011-12-21T11:42:56 that is only those still in the finals that made the first cut 2011-12-21T11:43:57 *** dapplegate has joined #aichallenge 2011-12-21T11:44:07 *** Apophis has quit IRC (Disconnected by services) 2011-12-21T11:44:30 *** Apophis_ has joined #aichallenge 2011-12-21T11:44:37 oh wow, a (very) few people have in the 30s and 40s. i guess that is the algorithm at work. 2011-12-21T11:44:40 you will notice mr. 88 games is gone 2011-12-21T11:45:11 *** dapplegate is now known as danapplegate 2011-12-21T11:45:30 McLeopold: so there is huge peak at 13 games, should it be? 2011-12-21T11:45:49 *** Zedenstein has joined #aichallenge 2011-12-21T11:46:00 oh 2011-12-21T11:46:02 bad query 2011-12-21T11:46:10 that is everyone who didn't make the cut :) 2011-12-21T11:46:16 heh 2011-12-21T11:46:28 lol 2011-12-21T11:46:43 Boo. My first timeout in finals (a turn 1 one) 2011-12-21T11:46:57 http://pastebin.com/icbWGMwg 2011-12-21T11:47:17 hi mr 88 2011-12-21T11:47:20 holy cow, pairing cutoff 5000 and 37 servers 2011-12-21T11:47:50 wow, no wonder why I am the seed 2011-12-21T11:48:09 I'm at 25 games 2011-12-21T11:48:25 looks better, but still too large dispersion 2011-12-21T11:48:29 McLeopold: DROP TABLE 2011-12-21T11:48:35 which one? 2011-12-21T11:48:37 (and just played 4 games) 2011-12-21T11:48:56 DROP ALL THE TABLES 2011-12-21T11:49:18 we all know that's how *you* manage a server 2011-12-21T11:49:39 ;) 2011-12-21T11:49:49 :) 2011-12-21T11:50:48 somehow http://images.memegenerator.net/instances/280x280/9728960.jpg wasnt the #1 google image search hit 2011-12-21T11:50:55 i hope this manipulation wont affect trueskill's fairness 2011-12-21T11:51:02 88 2011-12-21T11:51:02 better report that to search quality 2011-12-21T11:51:32 well 88 messes up if you use "score" to rank instead of "mu" 2011-12-21T11:51:37 but ideally mu is still usable 2011-12-21T11:51:42 *** moxian has quit IRC (Read error: Connection reset by peer) 2011-12-21T11:52:07 we still have quite a few starter bots in there 2011-12-21T11:52:15 yeah i bet 2011-12-21T11:52:29 cutoff at 3000? 2011-12-21T11:52:30 with hills, starter bots can do better than hunter or lefty 2011-12-21T11:52:53 a1k0n: hahaha, exactly what i meant ;) 2011-12-21T11:52:55 a1k0n: http://aichallenge.org/server_stats.php 2011-12-21T11:53:12 *** moxian has joined #aichallenge 2011-12-21T11:53:33 avdg: yes thats what i was looking at 2011-12-21T11:53:40 what is the pairing cutoff? 2011-12-21T11:53:48 McLeopold, is it right, that at least one bot in each game should have minimal count of games? 2011-12-21T11:53:50 *** protocolocon has joined #aichallenge 2011-12-21T11:54:15 bearoff_w: yes 2011-12-21T11:54:22 conor_f: the number of bots not playing imo 2011-12-21T11:54:22 <|UncleVasya|> conor_f: bots with low rank do not play anymore 2011-12-21T11:54:30 *** erdman has joined #aichallenge 2011-12-21T11:54:30 Server #743 2011-12-21T11:54:31 0.0% fail 2011-12-21T11:54:36 Server #750 2011-12-21T11:54:36 0.0% fail 2011-12-21T11:54:39 McLeopold: I saw on lastest games, FE http://aichallenge.org/visualizer.php?game=306683 - there is 25 minimal count 2011-12-21T11:54:50 Accoun: looks like an advertisement ;-) 2011-12-21T11:54:51 number of bots which are playing i thought 2011-12-21T11:54:57 sorry, 18 2011-12-21T11:54:59 *** katebus has joined #aichallenge 2011-12-21T11:55:15 which doesnt matter 2011-12-21T11:55:16 inb4 someone whining about totaalnet servers 2011-12-21T11:55:16 a1k0n: it started at none 2011-12-21T11:55:24 only 2 good servers 2011-12-21T11:55:41 * avdg starts thinking 2011-12-21T11:55:47 <|UncleVasya|> Now only starter bots were cuted off. 2011-12-21T11:55:50 *** nha has quit IRC (Ping timeout: 248 seconds) 2011-12-21T11:56:30 bearoff_w: there was a 16 in there, meaning they had 15 2011-12-21T11:56:30 what does that % fail catually mean? 2011-12-21T11:57:12 Accoun: fail rate includes bots that timeout or crash 2011-12-21T11:57:20 McLeopold: sorry, my mistake 2011-12-21T11:57:47 it mean than all ozer servers is bad 2011-12-21T11:57:58 It means you don't know jack about statistics 2011-12-21T11:58:05 Accoun: your would be better, huh? 2011-12-21T11:58:20 Fluxid> http://tcpants.com/player/Accoun12301 let play 2011-12-21T11:58:22 ozer==other? 2011-12-21T11:58:38 *** ikaros has joined #aichallenge 2011-12-21T11:58:38 ozer==your? 2011-12-21T11:58:45 *** denes has quit IRC () 2011-12-21T11:58:48 ozer=other 2011-12-21T11:58:50 a1k0n, avdg: I'm almost sure it's the number of player still playing. It went from 9k towards 5k. And if you look at current games, no one is further than 5k 2011-12-21T11:58:56 Accoun: i let you play 2011-12-21T11:59:03 it u grinvich gargon 2011-12-21T11:59:07 in u grinvich gargon 2011-12-21T11:59:14 *** codetiger has quit IRC () 2011-12-21T11:59:15 Surya: yeah, I was thinking (which means mostly something not good) 2011-12-21T11:59:22 in your greenwich jargon? 2011-12-21T11:59:27 yes, pairing cutoff is the rank, so only 5000 in the finals 2011-12-21T11:59:38 I am pretty sure that was rot13 2011-12-21T11:59:52 tcpants is uber slow, pages load for hours :| 2011-12-21T12:00:02 *** raemde_ has joined #aichallenge 2011-12-21T12:00:03 at least for me 2011-12-21T12:00:31 <|UncleVasya|> Fluxid: you know what people will answer you, right? :) 2011-12-21T12:00:44 that my server was uber slow too? 2011-12-21T12:01:02 <|UncleVasya|> No. 2011-12-21T12:01:14 <|UncleVasya|> Something like 'So bring your server back!' 2011-12-21T12:01:20 http://tcpants.com/player/Accoun12301 this server? 2011-12-21T12:01:25 |UncleVasya|: please, no 2011-12-21T12:01:29 Accoun: yes 2011-12-21T12:01:43 someone run many copies taimout in st move bots 2011-12-21T12:01:49 someone run many copies taimout in 1st move bots 2011-12-21T12:01:58 ban it please 2011-12-21T12:02:06 Accoun: i still have hard time understanding why didn't you submit your bot. you say something about totaalnet and canadians but it makes little sense for me 2011-12-21T12:02:31 wow, randomness scared me this time, I was scared to loose a hill in http://aichallenge.org/visualizer.php?game=306590&user=280 2011-12-21T12:02:32 *** raemde has quit IRC (Ping timeout: 252 seconds) 2011-12-21T12:02:36 <|UncleVasya|> Fluxid: I do not ask about it. You might remember I can't use any tcp-server because of my crapy connection 2011-12-21T12:02:39 it not worck on totalnet servers 2011-12-21T12:02:50 al replase it with starter 2011-12-21T12:02:56 im replase it with starter 2011-12-21T12:03:06 uh 2011-12-21T12:03:14 (well, the one on the bottom) 2011-12-21T12:03:15 no tame to make special version 2011-12-21T12:03:15 aichallenge: McLeopold epsilon * r631976a / sql/2_generate_matchup.sql : fix for possible matchup race condition - http://git.io/852X9w 2011-12-21T12:03:22 no time to make special version 2011-12-21T12:03:48 Accoun: if it works with tcp servers then it would work on official server 2011-12-21T12:03:57 http://tcpants.com/player/Accoun12301 here i play 2011-12-21T12:04:03 not necessarily 2011-12-21T12:04:19 http://tcpants.com/player/Accoun12301 if be more oponents i can make version to beat a1con 2011-12-21T12:04:58 *** praveen__ has quit IRC (Quit: Page closed) 2011-12-21T12:04:58 on all maps 2011-12-21T12:05:17 *** bearoff_w has left #aichallenge 2011-12-21T12:05:37 *** bearoff_w has joined #aichallenge 2011-12-21T12:06:06 Server #750 and Server #743 only ok all other give 5-13 % fail 2011-12-21T12:06:13 low prfomanse 2011-12-21T12:07:09 low memory or CPU speed 2011-12-21T12:07:16 Please stop 2011-12-21T12:07:34 You've been on this server-rant for days now 2011-12-21T12:07:47 Do we know when will the next cutoff happen? I can see that there are still some starter pack in the top 5000 : http://aichallenge.org/profile.php?user=12101 2011-12-21T12:07:53 Accoun: 17:57:12 < McLeopold> Accoun: fail rate includes bots that timeout or crash 2011-12-21T12:07:53 pairofdice: whoo, you saved me I almost fed the troll :( 2011-12-21T12:08:01 days? three whole contests 2011-12-21T12:08:17 Accoun: and you still make little sense. that was part of challenge: use little time and little memory 2011-12-21T12:08:41 *** sigh has joined #aichallenge 2011-12-21T12:08:48 today is much better than yesterdeay 2011-12-21T12:08:53 at least twice as fast 2011-12-21T12:08:58 pairofdice> i try to post in last 3 day, it radicaly slower than desctop, no time to make especial version 2011-12-21T12:09:06 and after another cut-off it will be like 4 times 2011-12-21T12:09:57 Accoun: i'm sorry you have cluster of 9001 cold cpu cores 2011-12-21T12:10:14 Fluxid> is 2 servers all ok all oser bad-> problems not with bots problrm with servers 2011-12-21T12:10:15 *** Murashka has joined #aichallenge 2011-12-21T12:10:18 *** Speedy_Consoles has quit IRC (Quit: KVIrc 4.0.1 Insomnia http://www.kvirc.net/) 2011-12-21T12:10:38 *** amstan has joined #aichallenge 2011-12-21T12:10:38 *** ChanServ sets mode: +o amstan 2011-12-21T12:10:50 we dont know how many games each server ran 2011-12-21T12:10:55 ¯\(°_o)/¯ 2011-12-21T12:11:10 *** Blkt has quit IRC (Quit: ERC Version 5.3 (IRC client for Emacs)) 2011-12-21T12:11:18 accoun how it is problem with servers 2011-12-21T12:11:26 you have 500 ms limit 2011-12-21T12:11:33 and i believe everyone is using that limit 2011-12-21T12:11:45 low memory or CPU speed 2011-12-21T12:11:54 Server #750 and Server #743 only ok all other give 5-13 % fail 2011-12-21T12:12:10 so it is 3 minutes per user per cpu core 2011-12-21T12:12:24 well, if the bot timeouts then it is fail 2011-12-21T12:12:31 it have 10 cores? 2011-12-21T12:12:32 just look like GreedyBot always timeouts 2011-12-21T12:12:54 there are many weak timeouting bots there 2011-12-21T12:12:56 Accoun: somehow i timeouted only once, and xathis never. 2011-12-21T12:13:05 and of course - sometimes there will be a server problem 2011-12-21T12:13:17 I have never timeouted 2011-12-21T12:13:36 I have check and 30 ms before limit I am out finishing the turn 2011-12-21T12:13:38 9% with all bots and 99% with some bots 2011-12-21T12:13:50 * Fluxid gives up 2011-12-21T12:14:05 thers same bots may be very god 2011-12-21T12:14:11 there same bots may be very god 2011-12-21T12:14:39 * janzert likes how 0 becomes 9% 2011-12-21T12:14:39 ok, remember thath some garbage collected languages may have problem, when someone generates many objects per turn 2011-12-21T12:14:42 like 500MB 2011-12-21T12:14:44 *** GreenTea has quit IRC (Quit: Page closed) 2011-12-21T12:15:12 *** bretep has joined #aichallenge 2011-12-21T12:15:36 *** ert has joined #aichallenge 2011-12-21T12:16:21 now 743 has 3% fail 2011-12-21T12:16:26 *** mega1 has joined #aichallenge 2011-12-21T12:16:27 i dont now but in fact game servers radicaly slow than desctop 2011-12-21T12:16:31 so it couldnt have had many games at 0 2011-12-21T12:16:44 it the reson why i don post my bot 2011-12-21T12:16:55 no time to make special version 2011-12-21T12:17:01 I'm lucky I'm reusing my objects, but I still don't know how much garbage there is left on my bot (with the gc) 2011-12-21T12:17:01 that is true 2011-12-21T12:17:13 but that is the constraint everyone is under 2011-12-21T12:17:46 if it be 1 core por bot it will be ok for me 2011-12-21T12:17:50 all the worker stats on the server stats page are for the previous half hour 2011-12-21T12:17:52 if it be 1 core per bot it will be ok for me 2011-12-21T12:17:57 Accoun: it is 2011-12-21T12:17:59 Accoun maybe you have super fast desktop ;-) 2011-12-21T12:18:18 5Y old 2011-12-21T12:18:44 intel core duo overcloked to 3000GZ 2011-12-21T12:18:47 maybe you had different compiler runner whatever? 2011-12-21T12:18:48 Accoun: EVERYONE has to put of with 500ms turntime 2011-12-21T12:18:48 *** TheLinker has joined #aichallenge 2011-12-21T12:18:50 everyone 2011-12-21T12:19:01 are you timing your simplex solver to make sure it runs <500ms? 2011-12-21T12:19:01 i dont now but in fact game servers radicaly slow than desctop 2011-12-21T12:19:07 tcp server timeout is 5s 2011-12-21T12:19:10 And multithreading is prohibited 2011-12-21T12:19:18 Servers seems to have quite a variable turntime 2011-12-21T12:19:18 it the reson why i don post my bot, no time to make especial version 2011-12-21T12:19:19 I was thinking I have fast one ;-) laptop core duo at 1.8 GHz 2011-12-21T12:19:19 *** iglo has joined #aichallenge 2011-12-21T12:19:23 Accoun, that shouldn't matter at all 2011-12-21T12:19:36 katebus: thats as fast as my previous one 2011-12-21T12:19:48 sad its screen got broken 2011-12-21T12:19:53 Sorry to annoy you guys with the same question over and over but: When is the next cutoff? (approx) 2011-12-21T12:20:15 also did you even try it? 2011-12-21T12:20:27 Surya: we don't know, we look at results, then decide in the moment 2011-12-21T12:20:43 McLeopold: ok, thank you :) 2011-12-21T12:21:10 bye bye, kids want to play with me ;-) there is reality besides the ants contest 2011-12-21T12:21:17 *** Surya has quit IRC (Quit: Page closed) 2011-12-21T12:22:24 *** xathis has joined #aichallenge 2011-12-21T12:22:24 i cant make any beter 2011-12-21T12:22:26 *** katebus has quit IRC (Quit: Page closed) 2011-12-21T12:22:29 i cant make any better 2011-12-21T12:22:53 5 games per day dont give me test well stability 2011-12-21T12:23:33 *** onion42 has quit IRC (Quit: Page closed) 2011-12-21T12:23:46 in 1st tron contest game servers be faster than desctop, i hope on it 2011-12-21T12:24:01 * avdg wonders how many games xathis need to reach mu 100 2011-12-21T12:25:02 i make time control to .36 sec + some fixetd time algoritmas than take .016 sec on my desctop 2011-12-21T12:25:29 play some games ok than start to tamouts an crash 2011-12-21T12:25:39 my be memory priblems 2011-12-21T12:26:27 i dont now 2011-12-21T12:26:35 my be canadian sbotage 2011-12-21T12:26:42 my be canadian sabotage 2011-12-21T12:27:22 in final i replase it with starter 2011-12-21T12:27:23 that is the most likely explanation 2011-12-21T12:27:28 Are they still being sneaky? :-) 2011-12-21T12:27:37 heeeeeey mega1! 2011-12-21T12:27:48 you joined just in time for another Accoun rant, how opportune 2011-12-21T12:27:49 hello 2011-12-21T12:28:13 I figured that's listening in at this point is not too risky. 2011-12-21T12:28:17 rant? 2011-12-21T12:28:37 *** thagomizr has quit IRC (Quit: thagomizr) 2011-12-21T12:29:04 http://tcpants.com/player/Accoun12301 here my bot worck well 2011-12-21T12:29:17 has it been a good challenge? 2011-12-21T12:29:31 bad for me 2011-12-21T12:29:57 mega1: nice to see you! 2011-12-21T12:29:59 pretty interesting. my bot does pretty well for how simple it is 2011-12-21T12:30:11 Accoun: timing out? 2011-12-21T12:30:19 yes 2011-12-21T12:30:52 Is it known how top bots work? Or how different they are? 2011-12-21T12:31:08 *** conor_f has quit IRC (Ping timeout: 252 seconds) 2011-12-21T12:31:20 code has been posted. approaches are similar 2011-12-21T12:31:29 mine's different. :) 2011-12-21T12:31:54 i decided to do some monte carlo stuff this time 2011-12-21T12:31:56 *** conor_f has joined #aichallenge 2011-12-21T12:32:03 everyone else went for diffusion? 2011-12-21T12:32:19 no, nobody uses diffusion at the top 2011-12-21T12:32:23 mc looks more promising this time 2011-12-21T12:32:25 all use simples solve for matrix game? 2011-12-21T12:32:27 unless you count bfs as diffusion 2011-12-21T12:32:43 all use simplex solve for matrix game? 2011-12-21T12:32:54 no, Accoun i think yours is unique there 2011-12-21T12:33:15 some gots have name Simplex 2011-12-21T12:33:17 *** teapotahedronfan has joined #aichallenge 2011-12-21T12:33:19 some bots have name Simplex 2011-12-21T12:33:22 the multi source BFS is kind of like diffusion, I guess. 2011-12-21T12:33:35 *** hadron is now known as teapotahedron_ko 2011-12-21T12:33:36 that is all I used, besides some combat code 2011-12-21T12:33:39 greentea uses brute force for small clumps of ants, xathis too i think 2011-12-21T12:33:46 *** keith__ has quit IRC (Ping timeout: 258 seconds) 2011-12-21T12:33:51 *** qwebirc964812 has joined #aichallenge 2011-12-21T12:34:08 brute force for dont die? 2011-12-21T12:34:26 *** qwebirc964812 is now known as teapotahedron_ni 2011-12-21T12:34:49 well, one-step minimax 2011-12-21T12:35:10 why no one want play with me on http://tcpants.com/player/Accoun12301 ? 2011-12-21T12:35:12 @rankings 2011-12-21T12:35:14 teapotahedronfan: Top 10 players: xathis(87.9), teapotahedron(83.7), meduza(83.1), FlagCapper(82.7), BenJackson(81.5), _Atsky_(81.4), bix0r4ever(81.1), Komaki(81.0), Speedy_Consoles(81.0), cheeser(81.0) 2011-12-21T12:35:24 *** moxian is now known as teapotahedron_mo 2011-12-21T12:35:26 only a1k0n from hy reited 2011-12-21T12:35:28 teapotahedron --- the best! 2011-12-21T12:35:29 only a1k0n from hi reited 2011-12-21T12:35:38 teapotahedronfan: welcome to the club 2011-12-21T12:35:40 bro 2011-12-21T12:35:49 hello mega1 from hungary :D 2011-12-21T12:35:51 heh 2011-12-21T12:35:54 so it's domain specific tweaking again as opposed to silver bullets 2011-12-21T12:36:00 yep 2011-12-21T12:36:19 has the next game been decided on yet? 2011-12-21T12:36:22 was MC a development time saver? 2011-12-21T12:36:25 teapotahedronfan, thanks for the support 2011-12-21T12:36:29 mine was a fairly general solution but since coordination between ants is necessary it tends to suicide once in a while 2011-12-21T12:36:33 mega1: hi 2011-12-21T12:36:37 anyone classifies enemies learning from then? 2011-12-21T12:36:46 amstan, dmj111: hi guys 2011-12-21T12:36:55 mega1: yeah, i didn't spent much time on the bot 2011-12-21T12:37:02 that is why I limited my participation time this time... I only want silver bullets now :) 2011-12-21T12:37:29 working through the particulars of the attacking influence of each ant deterministically would have been a better approach 2011-12-21T12:37:36 a1k0n: your reank is pretty impressive then 2011-12-21T12:37:50 but it seems that efficient foraging is really what's important, as long as you have some semblance of combat and defense 2011-12-21T12:37:51 *** teapotahedron_mo has quit IRC (Read error: Connection reset by peer) 2011-12-21T12:37:59 mega1: no one is more surprised than me 2011-12-21T12:37:59 I only saw GreenTea's jar though he said his source was available aswell 2011-12-21T12:38:12 i tried to do a similar thing for planetwars and it got me nowhere 2011-12-21T12:39:14 mega1: the rumor was that your wife forbid you to do this contest? 2011-12-21T12:39:31 *** mega1` has joined #aichallenge 2011-12-21T12:39:31 *** cofesor has quit IRC (Ping timeout: 258 seconds) 2011-12-21T12:39:56 a1k0n: heh, it was mutual agreement :-) 2011-12-21T12:40:01 oops 2011-12-21T12:40:06 *** mega1` has left #aichallenge 2011-12-21T12:40:22 so no decision on the next game then? 2011-12-21T12:40:31 bbl 2011-12-21T12:40:37 we just had another baby so i was on paternity leave for most of this contest. after i went back to work i left my bot alone, haha 2011-12-21T12:41:10 congrats for the baby as well 2011-12-21T12:41:15 thanks 2011-12-21T12:41:26 be back after dinner 2011-12-21T12:41:28 yes , congrats. 2011-12-21T12:42:14 Accoun: i could beat you several times on tcp 2011-12-21T12:42:14 *** teapotahedronfan has left #aichallenge 2011-12-21T12:42:27 <|UncleVasya|> wait, wasn't this guy a bocsimacko owner? 2011-12-21T12:42:32 yep 2011-12-21T12:42:35 lol 2011-12-21T12:42:37 yEP 2011-12-21T12:42:42 the reason a picked up lisp 2011-12-21T12:42:43 a1k0n: congrats a1k0n! so 2 contests and 2 babies :-) 2011-12-21T12:42:45 a->i 2011-12-21T12:42:52 indeed, haha 2011-12-21T12:43:03 3 contests 2011-12-21T12:43:12 2 babies not so many 2011-12-21T12:43:16 <|UncleVasya|> heh, I am in one room with Mr. Melis! Can you give me an autograph? :D 2011-12-21T12:43:25 2 is plenty for sure 2011-12-21T12:43:34 not 2011-12-21T12:43:39 yeah i did have a crappy entry for the planetwars contest 2011-12-21T12:43:48 take ne yang waife 2011-12-21T12:43:59 old woman cant make good babies 2011-12-21T12:44:00 heh. i can't afford to have any more babies in california 2011-12-21T12:44:10 it's not a question of biology. 2011-12-21T12:44:32 how problems? 2011-12-21T12:44:42 smal house? 2011-12-21T12:44:56 small, expensive house, and day care and school is very expensive 2011-12-21T12:45:29 by new one 2011-12-21T12:45:33 bue new one 2011-12-21T12:45:39 bye new one 2011-12-21T12:45:40 heh. we can't even sell our old one in wisconsin (which is huge) 2011-12-21T12:45:47 how much does a funeral service cost there? 2011-12-21T12:45:53 no idea 2011-12-21T12:46:04 *** praveen___ has joined #aichallenge 2011-12-21T12:46:14 go live in wisconsin 2011-12-21T12:46:25 perhaps i should 2011-12-21T12:46:27 aichallenge: McLeopold epsilon * rcf7c1de / website/header.php : add countdown desc to server_info - http://git.io/9_LhCA 2011-12-21T12:47:10 *** Antimony has joined #aichallenge 2011-12-21T12:47:47 *** thagomizr has joined #aichallenge 2011-12-21T12:48:27 @rankings 30 2011-12-21T12:48:28 Ashoka: Top 30 players: xathis(87.9), teapotahedron(83.7), meduza(83.1), FlagCapper(82.7), BenJackson(81.5), _Atsky_(81.4), bix0r4ever(81.1), Komaki(81.0), Speedy_Consoles(81.0), cheeser(81.0), runevision(80.7), Memetix(80.5), delineate(80.4), JG.WAS(79.7), GreenTea(79.6), protocolocon(79.5), oldman(79.5), lazarant(79.4), a1k0n(79.2), strcat(79.1), bistrik(78.7), pguillory(78.6), fourmidable(78.4), (1 more message) 2011-12-21T12:48:33 *** teapotahedron_ko is now known as tetrahedron 2011-12-21T12:49:25 aichallenge: McLeopold epsilon * r4e4705d / website/header.php : add countdown desc to server_info - http://git.io/jufvkw 2011-12-21T12:49:40 on TCP i be beter than strcat 2011-12-21T12:49:49 *** cyphase has quit IRC (Ping timeout: 240 seconds) 2011-12-21T12:52:10 aichallenge: McLeopold epsilon * r169ae60 / website/header.php : add countdown desc to server_info - http://git.io/PoI7Iw 2011-12-21T12:52:11 Haha, just noticed: on main page, top left: "Timer goes here" 2011-12-21T12:52:53 im possible 2011-12-21T12:52:56 Aaand now it's gone 2011-12-21T12:53:01 i could beat you sometimes 2011-12-21T12:53:10 and im sure i could not have beaten strcat 2011-12-21T12:53:11 *** Antimony_ has joined #aichallenge 2011-12-21T12:53:19 horaay, I'm back up to 45th 2011-12-21T12:54:02 Yay, finals extended :) 2011-12-21T12:54:10 mcstar> http://ants.fluxid.pl/search?name=GyG 2011-12-21T12:54:11 by one day 2 hours 2011-12-21T12:54:14 *** Antimony has quit IRC (Ping timeout: 252 seconds) 2011-12-21T12:54:21 *** Antimony_ is now known as Antimony 2011-12-21T12:54:49 Accoun: what do you want with that? 2011-12-21T12:54:54 i dont understand 2011-12-21T12:55:08 mcstar> http://ants.fluxid.pl/search?name=GyG last days TSP be overloaded, alot of timouts 2011-12-21T12:55:20 mcstar> http://ants.fluxid.pl/search?name=GyG last days TSP be overloaded, alot of timeouts 2011-12-21T12:55:40 traveling salesman problem? 2011-12-21T12:55:54 joke 2011-12-21T12:56:03 will there be further cutoffs? 2011-12-21T12:56:32 ikaros: as a matter of fact... 2011-12-21T12:56:38 Accoun: why couldnt you beat me there? http://tcpants.com/replay.19397 2011-12-21T12:56:43 cutoff in motion again, heading to 3000 over the next 2 hours 2011-12-21T12:56:58 oh ok =) 2011-12-21T12:57:16 it will probably stay there for longer than it did at 5000 2011-12-21T12:57:43 mcstar> need mor moves 2011-12-21T12:57:50 mcstar> need more moves 2011-12-21T12:57:53 you have a plan on how many cutoffs there will be and how many bots will be in the last list 2011-12-21T12:58:14 mcstar> lusind some hils it random 2011-12-21T12:58:25 *** Scryer has joined #aichallenge 2011-12-21T12:58:31 Accoun: HOW much time does your bot use on average? 2011-12-21T12:59:25 *** sigh has quit IRC (Remote host closed the connection) 2011-12-21T12:59:27 *** protocolocon has quit IRC (Ping timeout: 258 seconds) 2011-12-21T12:59:32 mcstar> .36 target time 2011-12-21T13:00:12 mcstar> .36 + some alogitmas with .015 max time 2011-12-21T13:00:47 *** amstan_ has joined #aichallenge 2011-12-21T13:00:47 *** ChanServ sets mode: +o amstan_ 2011-12-21T13:01:22 * avdg wonders if we need the counter for winner announcement 2011-12-21T13:01:39 ikaros: the target is stable top bot rankings by the end of the finals without taking the cutoff below 500 and giving everyone as many games as possible 2011-12-21T13:01:51 here http://ants.fluxid.pl/player/GyG12206 my bot, but alot timouts dont give good ranking 2011-12-21T13:02:09 4908 pairing cutoff 2011-12-21T13:02:10 * Scryer thinks the counter probably answers a FAQ that would otherwise soak up lots of organizer time. 2011-12-21T13:02:14 but the cutoff may go as low as 250 if need be 2011-12-21T13:02:34 Scryer: true 2011-12-21T13:03:50 *** thagomizr has quit IRC (Quit: thagomizr) 2011-12-21T13:03:52 hmm, I wonder at which position the starter bots are 2011-12-21T13:04:08 *** amstan has quit IRC (Ping timeout: 240 seconds) 2011-12-21T13:05:17 *** rlanday has joined #aichallenge 2011-12-21T13:05:32 avdg: no higher than 2600, looking at the "Won at 1" games on "Latest Games." 2011-12-21T13:06:05 The "Won at 1" games have dropped from 34% before the initial cutoff to 23% now. 2011-12-21T13:06:15 Accoun: M was a good opponent, you should measure your bot according to his 2011-12-21T13:06:19 MXX 2011-12-21T13:07:55 *** kilae has joined #aichallenge 2011-12-21T13:08:03 mcstar> need to fix map set/map generator and i will tyne up for it 2011-12-21T13:08:55 *** mj41 has quit IRC (Ping timeout: 255 seconds) 2011-12-21T13:11:16 *** katebus has joined #aichallenge 2011-12-21T13:11:30 *** GreenTea has joined #aichallenge 2011-12-21T13:11:41 Is a1k0n on tcpants the same version as on the server? 2011-12-21T13:12:35 or better 2011-12-21T13:12:39 probably the same 2011-12-21T13:13:18 bretep: same 2011-12-21T13:13:27 it's probably running on a faster machine 2011-12-21T13:13:42 see cutoff is decreased.. what is next stop? 2011-12-21T13:13:51 *** choas has joined #aichallenge 2011-12-21T13:13:57 GreenTea: 3000 2011-12-21T13:14:04 good 2011-12-21T13:14:10 *** TheLinker has quit IRC (Quit: Bye) 2011-12-21T13:15:03 Moving towards a cutoff of 3000 2011-12-21T13:15:03 *** praveen___ has quit IRC (Quit: Page closed) 2011-12-21T13:15:10 hehe new timer 2011-12-21T13:15:15 in about 2 hours 2011-12-21T13:15:47 my bot is really stinking it up on the contest site 2011-12-21T13:16:15 wow, I beat account?? 2011-12-21T13:16:35 oh, timeout 2011-12-21T13:16:46 yeah... apparently he didn't solve them at all 2011-12-21T13:16:47 I had a draw against, a1k0n on tcpants. :v 2011-12-21T13:17:04 *** thagomizr has joined #aichallenge 2011-12-21T13:17:06 McLeopold> need more games 2011-12-21T13:17:22 McLeopold> compyter be swithet off 2011-12-21T13:17:27 Accoun: I never finished by bot 2011-12-21T13:17:36 heh, that map is too big for a 1v1 2011-12-21T13:17:58 yeah, the tcp guys didn't really keep up on official maps 2011-12-21T13:19:01 http://tcpants.com/replay.19474 lol 2011-12-21T13:19:18 *** itzkow has joined #aichallenge 2011-12-21T13:20:39 good defense, need more offense 2011-12-21T13:21:25 <|UncleVasya|> Accoun: Have you released your PW-bot source code? 2011-12-21T13:21:32 Does tcpants keep the old threads? 2011-12-21T13:21:53 Just hit 100% cpu and see several python threads running :x 2011-12-21T13:22:22 I think Accoun just runs giant server farms doing neural network tuning, which is why is is always 2nd or 3rd in ranks 2011-12-21T13:23:30 a1k0n> need tune up 2011-12-21T13:24:25 McLeopold> neural network suck ) 2011-12-21T13:24:59 yeah, they are the 2nd best way to do anything :) 2011-12-21T13:25:24 *** bluegaspode has joined #aichallenge 2011-12-21T13:26:02 Accoun: ah, so that was you! 2011-12-21T13:26:19 *** Surya has joined #aichallenge 2011-12-21T13:26:55 anyone have a good algorythm for symmetry detection? 2011-12-21T13:27:15 http://aichallenge.org/visualizer.php?game=307613&user=757 awesome 2011-12-21T13:27:44 find all possible symmetries, eliminate them as you acquire counterexamples? 2011-12-21T13:28:00 Trapped 2011-12-21T13:28:13 for each square you reveal, compare water/not water against what you know with currently possible symmetries 2011-12-21T13:28:26 Fluxid: hahah 2011-12-21T13:28:31 McLeopold did you see my writeup... https://github.com/jcdny/bugnuts/wiki/Symmetry 2011-12-21T13:28:34 *** cyphase has joined #aichallenge 2011-12-21T13:28:49 a1k0n: yeah, I can describe it, coding it is harder 2011-12-21T13:28:50 Fluxid, filled like a sand) 2011-12-21T13:28:52 jcdny: wow nice 2011-12-21T13:29:20 for multihill maps I have the correct symmetry typically before turn 10. 2011-12-21T13:29:49 that would have really helped my combat 2011-12-21T13:30:33 parts of my symmetry code are reasonably elegant then there is a big slug of crap taco that I meant to clean up but didn't 2011-12-21T13:30:52 that's very clever 2011-12-21T13:31:21 sadly I was appreciably less clever on my combat which is why I languish at 160th. 2011-12-21T13:31:40 what an odd number to timeout at in multiple matches: http://forums.aichallenge.org/viewtopic.php?f=27&t=2198 2011-12-21T13:32:36 a1k0n: also, you can look for food spawned on the same turn 2011-12-21T13:32:59 I was trying to use that to find possible enemy hill locations 2011-12-21T13:33:13 right 2011-12-21T13:33:22 but finding an example takes quite a few turns 2011-12-21T13:33:24 It does that. Another thing is with symmetry you know food spawns you can't see 2011-12-21T13:33:42 yeah, so you see one food, you can assume the other places 2011-12-21T13:33:48 then you need less map coverage 2011-12-21T13:34:14 as long as you know your territory is secure 2011-12-21T13:34:25 but I think using antimatroids method can help with that 2011-12-21T13:36:02 The biggest problem with my map symmetry code is that I then have too many hill targets and I don't prioritize them - I just start marching off to all the ones I know about. 2011-12-21T13:36:17 that is what everyone who tried symmetry detection said 2011-12-21T13:37:22 Bad handling of hills 2011-12-21T13:37:25 Not of symmetry 2011-12-21T13:37:54 I gained a good number of skill points when my ninja code started working well -- every hill I know about gets one risk-averse ninja. 2011-12-21T13:37:58 shuffle(hills) for hill in hills: prime_target = hill 2011-12-21T13:38:19 Doesn't do anything against good bots, ofc. 2011-12-21T13:38:34 yeah I ended up in about the same place although I sent a pair if available... 2011-12-21T13:38:57 ninja ants are awesome 2011-12-21T13:39:04 Even with good bots in the game, though, it's helpful when trying to beat them to the weak ones. 2011-12-21T13:39:09 jcdny: the bit rotations are a bit unclear in your description 2011-12-21T13:39:19 ok will attempt to clarify 2011-12-21T13:39:27 tyvm for the feedback 2011-12-21T13:39:30 where do you put your bit positions for your example? 2011-12-21T13:39:56 you just read them off in order 2011-12-21T13:40:05 2<<(pos-1)? 2011-12-21T13:40:13 from the 0 rotation 2011-12-21T13:40:14 ? 2011-12-21T13:40:37 oh i see it now 2011-12-21T13:40:50 http://aichallenge.org/visualizer.php?game=307721&user=12 i cant believe i had 500 ants in the hill 2011-12-21T13:40:58 so for the 3x3 0th it's 123456789 and for the 90 degree its 741852963 2011-12-21T13:41:08 yeah 2011-12-21T13:41:08 a1k0n: do you have code that makes sure you emit an ent every turn? 2011-12-21T13:41:26 ant* 2011-12-21T13:41:34 mcstar: i setitimer() and break out of my sampler loop and then always generate greedy moves afterward 2011-12-21T13:44:18 *** rlanday has quit IRC (Ping timeout: 258 seconds) 2011-12-21T13:45:53 *** kire has joined #aichallenge 2011-12-21T13:46:59 *** conor_f has quit IRC (Ping timeout: 268 seconds) 2011-12-21T13:47:39 *** bluegaspode has quit IRC (Quit: Page closed) 2011-12-21T13:49:48 http://aichallenge.org/visualizer.php?game=307563&user=3473 2011-12-21T13:49:54 a bunch of these mazes are races 2011-12-21T13:49:56 http://forums.aichallenge.org/viewtopic.php?f=24&t=2210 2011-12-21T13:49:57 xathis and I lost 2011-12-21T13:50:24 BenJackson: no offense really, but its ncie 2011-12-21T13:50:59 it shows that ants dont really recognize their surroundings in time, and cant make a good global strategy always 2011-12-21T13:51:12 ants is very much a territory control game 2011-12-21T13:51:49 *** cyphase has quit IRC (Ping timeout: 244 seconds) 2011-12-21T13:52:06 thats an insane game 2011-12-21T13:52:17 and a lot of bots have the tendency to clump up and freeze instead of pushing through an opponent blockade of fewer ants 2011-12-21T13:52:49 komaki didnt stop single ants to fight, it let them forage 2011-12-21T13:54:04 BenJackson: any thoughts? 2011-12-21T13:54:30 it's not as extreme as some of the maps, but it does focus pretty closely on a single early event 2011-12-21T13:55:59 *** tomb___ has quit IRC (Quit: Page closed) 2011-12-21T13:56:00 http://aichallenge.org/visualizer.php?game=298890&user=8078 2011-12-21T13:56:04 no early event 2011-12-21T13:56:27 his territory control is interesting 2011-12-21T13:56:50 nobody really has territory control going there 2011-12-21T13:57:07 he spreads his ants very evenly 2011-12-21T13:57:11 thats what i meant 2011-12-21T13:57:25 it is a sort of control, cause depletes food 2011-12-21T13:57:33 *** jcdny has quit IRC (Quit: Leaving) 2011-12-21T13:57:35 and other dont get as many 2011-12-21T13:57:37 s 2011-12-21T13:58:38 @rankings 2011-12-21T13:58:39 dici: Top 10 players: xathis(86.9), FlagCapper(84.0), teapotahedron(83.7), meduza(82.6), Komaki(82.4), BenJackson(81.5), runevision(81.5), _Atsky_(81.4), bix0r4ever(81.1), Speedy_Consoles(81.0) 2011-12-21T13:58:52 *** Hexren has joined #aichallenge 2011-12-21T13:59:44 pguillory still not in top10 2011-12-21T14:00:18 nor chrish 2011-12-21T14:00:41 pguillory has had some bad losses, I think he doesn't have hill defence code 2011-12-21T14:01:06 *** mleise has joined #aichallenge 2011-12-21T14:02:40 I guess there were quite a bit of good uploads in the final days 2011-12-21T14:02:56 *** jcdny has joined #aichallenge 2011-12-21T14:03:37 McLeopold: in the game just linked here he does get ninjacapped 2011-12-21T14:03:43 but that random walk was a sea of mixed ants 2011-12-21T14:03:49 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-12-21T14:03:53 I'd like to hope I would have bottled all those guys up 2011-12-21T14:04:18 keeping 1 ant around might have helped 2011-12-21T14:05:16 *** Palmik has quit IRC (Remote host closed the connection) 2011-12-21T14:05:29 I treated my hills the same as an enemy: I always send in 1 ant + 2 per closer enemy 2011-12-21T14:05:32 *** Surya has quit IRC (Quit: Page closed) 2011-12-21T14:05:47 now, what I did with those ants sucked :) 2011-12-21T14:05:48 Didn't seem like it mattered to pguillory to have enemy-ants near his own hill. 2011-12-21T14:06:09 I suspect it doesn't matter as much on maze maps 2011-12-21T14:06:15 *** Hexren has left #aichallenge 2011-12-21T14:06:15 his hill wasnt in sight 2011-12-21T14:06:27 *** mviel_ has joined #aichallenge 2011-12-21T14:06:35 i made the mistake of keeping 1 ant near 1 hill always 2011-12-21T14:06:39 i.e. evrey hill 2011-12-21T14:06:43 I used to have a "get hill back in sight" code but normally now I don't lose track of it except on low food multi-hill 2011-12-21T14:06:48 the only condition is to have > 5 ants 2011-12-21T14:06:49 in which case it's hard to know if you need to gather or guard 2011-12-21T14:06:53 *** g0llum has joined #aichallenge 2011-12-21T14:06:55 i terribly lost a 6-hill map 2011-12-21T14:06:56 guarding forever is just going to lose eventually 2011-12-21T14:07:09 *** Antimony has joined #aichallenge 2011-12-21T14:07:33 *** Surya has joined #aichallenge 2011-12-21T14:09:35 *** mviel__ has quit IRC (Ping timeout: 252 seconds) 2011-12-21T14:10:09 *** Palmik has joined #aichallenge 2011-12-21T14:11:12 is McLeopold here? 2011-12-21T14:11:15 hi 2011-12-21T14:11:20 *** Palmik has quit IRC (Remote host closed the connection) 2011-12-21T14:11:35 If you still want symmetry code, you should take a look at my bot. I beleive I had the best symmetry detection of any bot 2011-12-21T14:12:02 okay, is it posted in the open source thread? 2011-12-21T14:12:48 Antimony: I played a game where map symmetry (not hill symmetry) would have helped me 2011-12-21T14:12:52 http://forums.aichallenge.org/viewtopic.php?f=24&t=2169 2011-12-21T14:12:54 one of those mazes with a solid wall 2011-12-21T14:13:03 my bot tried to path aruond the end of the wall (that it hadn't seen) 2011-12-21T14:13:11 what map? 2011-12-21T14:13:41 *** Accoun has quit IRC () 2011-12-21T14:13:54 *** QGazQ has joined #aichallenge 2011-12-21T14:15:34 https://github.com/jcdny/bugnuts/wiki/Symmetry 2011-12-21T14:15:40 How does he get the hashes 2011-12-21T14:16:22 Personally, I think the brute force approach is better. Hashes fail on maps with large numbers of identical subsections 2011-12-21T14:16:40 hashes never fail 2011-12-21T14:16:43 they slow down 2011-12-21T14:16:47 the brute force solution is too slow 2011-12-21T14:16:59 (should) 2011-12-21T14:17:26 does slow matter? you can spread the work out over the early turns 2011-12-21T14:17:29 brute force will always give you the correct answer 2011-12-21T14:17:32 thats what I do 2011-12-21T14:17:33 when you have few ans anyway 2011-12-21T14:17:44 *** cyphase has joined #aichallenge 2011-12-21T14:17:45 hashes will sometimes fail to detect a symmetry that you could have inferred 2011-12-21T14:18:10 It does not fail for any of the maps so far 2011-12-21T14:18:35 pick a replay and will see whose bot detects all teh symmetries fastest? 2011-12-21T14:18:53 *** McLeopold has left #aichallenge 2011-12-21T14:18:55 A few turns isn't going to make any difference 2011-12-21T14:19:06 And it leaves a lot more time for combat and other logic 2011-12-21T14:19:26 That are a fair bit more important than symmetry detection 2011-12-21T14:19:52 I do the detection at the end o the turn after I'm already doen with everything else 2011-12-21T14:20:53 *** |UncleVasya| <|UncleVasya|!kvirc@46-133-224-114.dialup.umc.net.ua> has quit IRC (Read error: Connection reset by peer) 2011-12-21T14:21:15 *** Ashoka has quit IRC (Quit: Page closed) 2011-12-21T14:21:37 Antimony: did your bot actually benefit from it?: 2011-12-21T14:21:38 *** Paradoxial has joined #aichallenge 2011-12-21T14:22:02 *** yoden has quit IRC (Ping timeout: 252 seconds) 2011-12-21T14:22:09 @rankings 2011-12-21T14:22:12 BenJackson: Top 10 players: xathis(87.5), FlagCapper(84.5), teapotahedron(83.7), meduza(82.6), Komaki(82.4), runevision(81.3), cheeser(81.2), _Atsky_(81.2), BenJackson(81.2), bix0r4ever(81.1) 2011-12-21T14:22:16 doh 2011-12-21T14:22:20 it's hard to tell 2011-12-21T14:22:42 but I'm guessing it did benefit 2011-12-21T14:22:48 just not very much 2011-12-21T14:23:04 hey at least you're in the top 10 2011-12-21T14:23:42 Antimony: for http://aichallenge.org/map.php?map=cell_maze/cell_maze_p04_23.map which symmetries do you recognize? 2011-12-21T14:23:58 sorry wrong example 2011-12-21T14:24:00 all of them :) 2011-12-21T14:24:34 *** katebus has quit IRC (Ping timeout: 258 seconds) 2011-12-21T14:24:35 BJ was first for quite some time 2011-12-21T14:24:38 http://aichallenge.org/map.php?map=cell_maze/cell_maze_p06_24.map 2011-12-21T14:25:03 instead of watching a game I must have lost I'm waiting until I go back up 2011-12-21T14:25:06 they've been coming in pairs 2011-12-21T14:25:07 it doesnt matter how intricate the water is 2011-12-21T14:25:08 you get both the translation and the mirror? 2011-12-21T14:25:11 it is still symmetrical 2011-12-21T14:25:13 yes 2011-12-21T14:25:25 I'm pretty sure my bot can handle any symmetry 2011-12-21T14:25:49 *** UncleVasya has joined #aichallenge 2011-12-21T14:26:10 the advantage of brute force is you always get correct results, and all of the results 2011-12-21T14:26:17 and you can even tell how many players there are 2011-12-21T14:26:56 Antimony: maybe not aperiodic tilings 2011-12-21T14:27:14 that only works with an infinte number of players 2011-12-21T14:27:23 and an infinite grid 2011-12-21T14:27:31 or wo wrapping 2011-12-21T14:28:10 if you give me an input file, I can run it and tell you approx how fast my bot would have found everything based on that input 2011-12-21T14:28:11 *** QGazQ|Work has joined #aichallenge 2011-12-21T14:28:22 Antimony is you code up anywhere? 2011-12-21T14:28:23 *** QGazQ has quit IRC (Quit: Page closed) 2011-12-21T14:28:28 *** QGazQ|Work is now known as QGazQ 2011-12-21T14:28:33 yes, I posted it on the forums 2011-12-21T14:28:42 cool I will take a look 2011-12-21T14:29:04 http://forums.aichallenge.org/viewtopic.php?f=24&t=2169 2011-12-21T14:29:21 i should print out some sources for the holidays to read while i wont have a computer 2011-12-21T14:29:24 *** xyxan has joined #aichallenge 2011-12-21T14:29:36 *** choas has quit IRC (Ping timeout: 240 seconds) 2011-12-21T14:29:42 my bot even has comments here and there 2011-12-21T14:29:58 guyz cutoff is now 3456 :P what is its limit? 2011-12-21T14:30:09 casino map - cell_maze_p04_22 just one mistake and someone gets all food and of course win 2011-12-21T14:30:51 xyxan: 3000 atm 2011-12-21T14:30:58 gr8 2011-12-21T14:31:11 *** lam has joined #aichallenge 2011-12-21T14:31:22 i wish 1000 will be a nice value or 1500 :P perhaps 2011-12-21T14:31:44 example http://aichallenge.org/visualizer.php?game=308244&user=589 2011-12-21T14:32:02 *** Accoun has joined #aichallenge 2011-12-21T14:32:02 *** gustavokambara has joined #aichallenge 2011-12-21T14:32:47 37 servers - wow! 2011-12-21T14:32:50 *** tomb___ has joined #aichallenge 2011-12-21T14:33:36 mcstar: will be mine among them? :) 2011-12-21T14:33:43 *** lama3oid has joined #aichallenge 2011-12-21T14:33:43 *** lam has quit IRC (Client Quit) 2011-12-21T14:33:49 *** Paradoxial has quit IRC (Ping timeout: 240 seconds) 2011-12-21T14:34:34 oh, finally i won some game 2011-12-21T14:34:37 Oh man, I'm slow 2011-12-21T14:34:57 http://aichallenge.org/visualizer.php?game=308252&user=757 very little satisfaction 2011-12-21T14:35:34 lol? 2011-12-21T14:35:58 UncleVasya: if i get my printer working 2011-12-21T14:36:15 *** valydo has joined #aichallenge 2011-12-21T14:36:36 mcstar: maybe you'd better wait for a commented version? 2011-12-21T14:36:40 Fluxid: looks very even except for the part where you defend your hill and he abandons his 2011-12-21T14:36:51 Fluxid: I would almost think that these bots are the same 2011-12-21T14:36:53 Not mch difference though 2011-12-21T14:37:02 apparently "Auto Referesh Plus" is a good Google Chrome Extension Lolx 2011-12-21T14:37:05 Win is where you find it, Fluxid. Declare success and win a harder one. :) 2011-12-21T14:37:27 BenJackson: that's why i'm not really satissfied. 2011-12-21T14:38:09 *** bearoff has joined #aichallenge 2011-12-21T14:39:31 ha, does the pairing cutoff slew to the new level over time? 2011-12-21T14:39:57 moving towards 3000 :D 2011-12-21T14:40:03 it slew several bots 2011-12-21T14:40:03 i would rather have an unexciting win, than a loss, especially if you have lots of time between games 2011-12-21T14:40:24 *** bmh has joined #aichallenge 2011-12-21T14:41:14 UncleVasya: if i print something ive got to do it tonite 2011-12-21T14:41:15 on the two play random walk maps, i found it difficult to balance exploration and food gathering with keeping the hill guarded 2011-12-21T14:41:15 http://tcpants.com/replay.19552 i love me! im so good 2011-12-21T14:41:44 Accoun: what's your bot's name on the official server? 2011-12-21T14:42:06 not again =) 2011-12-21T14:42:12 bmh: Timed out 2011-12-21T14:42:15 DoesntExist 2011-12-21T14:42:22 DoesntExist.ru 2011-12-21T14:42:24 ikaros: sorry I shouldn't troll :) 2011-12-21T14:42:35 russian bocks about game teory with with alot examples from war stategy 2011-12-21T14:42:44 *** Antimatroi has joined #aichallenge 2011-12-21T14:42:54 Damn my bot is sucking 2011-12-21T14:43:29 Antimatroi: don't worry, mine sucks as well 2011-12-21T14:44:39 and mine too ( have found 2 big bugs already ) 2011-12-21T14:44:49 Accoun is pacifistic 2011-12-21T14:44:59 Antimatroi: are you Anti Sprout? 2011-12-21T14:45:25 I'm guessing http://aichallenge.org/profile.php?user=98 2011-12-21T14:45:46 i know 2011-12-21T14:46:03 *** bmh has left #aichallenge 2011-12-21T14:46:17 *** yoden has joined #aichallenge 2011-12-21T14:46:40 he is a maximalist 2011-12-21T14:46:45 *** Redgis has joined #aichallenge 2011-12-21T14:48:58 Mine sucks much more than that, but since it ignores other players ants thats not really surprising 2011-12-21T14:50:30 :D 2011-12-21T14:50:39 does it ignore its own ants? 2011-12-21T14:50:50 what about food? 2011-12-21T14:50:55 If my bot can survive 200 turns, then its difficult to beat it :D, but i am getting suicidal attacks :( 2011-12-21T14:52:00 *** Regis has joined #aichallenge 2011-12-21T14:52:20 there should be bad food in finals. not worth to live 2011-12-21T14:52:36 expired food? 2011-12-21T14:52:49 something so 2011-12-21T14:52:57 tcpants seems down 2011-12-21T14:53:31 *** Regis has quit IRC (Client Quit) 2011-12-21T14:54:09 Nah it knows its own ants, and collects food, and attacks hills (thought not in a coordinated way) and explores. But not very advanced. As a plus I learnt some new java stuff 2011-12-21T14:54:45 some cloned my code :D 2011-12-21T14:54:48 one* 2011-12-21T14:55:27 * bretep cloning mcstars code to make him happy 2011-12-21T14:55:35 :O 2011-12-21T14:55:38 :O 2011-12-21T14:55:38 sould say I have the same code too ) 2011-12-21T14:56:00 im indeed happy :) 2011-12-21T14:56:05 lol 2011-12-21T14:56:13 it collects food, explores and so on, but NOT very advansed ( 2011-12-21T14:57:06 if i would have implemented combating :P i would be around 100 :D 2011-12-21T14:57:08 oh, as a minus, I learnt no any java stuff :( 2011-12-21T14:57:43 *** Antimatroi has quit IRC (Quit: Page closed) 2011-12-21T14:57:54 http://en.wikipedia.org/wiki/Skynet_(Terminator) 2011-12-21T14:58:52 Proposition for next comp, make the best Skynet 2011-12-21T14:59:22 :P 2011-12-21T14:59:41 *** Antimatroid has joined #aichallenge 2011-12-21T15:00:20 xkcd has thought about the Skynet possibility: http://xkcd.com/534/ 2011-12-21T15:00:32 Do we really expect to have results in just over 2 days? That seems very unlikely to me given how the top 10 currently looks 2011-12-21T15:00:44 need to ban AI contests - it will kill human race 2011-12-21T15:01:08 Antimatroid: I'll let you know in a sec 2011-12-21T15:01:10 I'm in a game 2011-12-21T15:01:12 Really, Antimatroid? I was gratified that everybody we expected to be in the top 10 is at least on page 1 already. 2011-12-21T15:01:18 at 9th I'm either going solidly back into the top 10 or out 2011-12-21T15:01:20 Antimony: how much time did you put into your symmetry code? 2011-12-21T15:01:33 Accoun: only cause we're all too busy writing stupid nots from addiction :D 2011-12-21T15:01:35 Antimatroid, now cutoff is 3000 and queue is much more fast 2011-12-21T15:01:39 *** jstemmer has quit IRC (Ping timeout: 255 seconds) 2011-12-21T15:01:58 *** valydo has quit IRC (Ping timeout: 248 seconds) 2011-12-21T15:02:07 *** Antimony has quit IRC (Ping timeout: 252 seconds) 2011-12-21T15:02:12 *** jstemmer has joined #aichallenge 2011-12-21T15:02:14 Mcstar what symmetry code? 2011-12-21T15:02:19 he is IMPOSSIBLE 2011-12-21T15:02:21 clearly we need an ai contest to write an ai to subdue the humans 2011-12-21T15:02:24 ;P 2011-12-21T15:02:26 times out every fing minute 2011-12-21T15:02:46 Antimatroid: http://forums.aichallenge.org/viewtopic.php?f=24&t=2169 2011-12-21T15:02:55 Much more fast? I've had 4 games since last night 2011-12-21T15:03:07 *** flowenol has joined #aichallenge 2011-12-21T15:03:23 Maybe everybody has to write 2 bots: one to be Skynet and one to be a small group of humans. 2011-12-21T15:03:43 Mcstar I was indicating that my bot has no symmetry code 2011-12-21T15:04:02 Antimatroid: sry? 2011-12-21T15:04:02 *** observer has joined #aichallenge 2011-12-21T15:04:15 who said you had? 2011-12-21T15:04:17 >> since last night - I don't know how long this period is ) 2011-12-21T15:04:20 :p 2011-12-21T15:04:23 you asked What symmtery? 2011-12-21T15:04:31 or semetery 2011-12-21T15:04:35 or whatever 2011-12-21T15:04:42 small group of humans ) 2011-12-21T15:04:53 may be cemetry 2011-12-21T15:04:57 With what operation? 2011-12-21T15:04:58 cementry 2011-12-21T15:05:04 does anyone implement cemetry? 2011-12-21T15:05:08 sedimentary 2011-12-21T15:05:10 I can't hear group without thinking mjaths 2011-12-21T15:05:15 the US needs to be in the top 10. go BenJackson 2011-12-21T15:05:24 whats mjaths? 2011-12-21T15:05:32 Norwegian maths. 2011-12-21T15:05:40 * Antimatroid is not good with this iPad, cut me some slack d: 2011-12-21T15:05:41 lol 2011-12-21T15:06:00 Antimatroid: i just dont understand your reaction, i was talking to antimony 2011-12-21T15:06:02 I thought I was bad watching my games on a kindle fire 2011-12-21T15:06:23 My games don't look that great on my smartphone. 2011-12-21T15:06:44 dont you have money to buy a effing desktop? 2011-12-21T15:06:48 my phone isnt powerful enough to render each turn 2011-12-21T15:07:01 Mcstar: now I'm with you, sorry morning :p 2011-12-21T15:07:14 Antimatroid: ok, it happends every day 2011-12-21T15:07:16 I can generally tell whats happening on mine, I check my games before I get out of bed in the morning 2011-12-21T15:07:17 -d 2011-12-21T15:07:19 My android phone does OK if I zoom it, but then I lose the big picture. 2011-12-21T15:07:36 *** Ashoka has joined #aichallenge 2011-12-21T15:07:47 Was that aimed at me? 2011-12-21T15:07:55 my android is under-powered or under-memoried. it is the lg ally. 2011-12-21T15:08:10 I'm away for Christmas and this is the easiest thing to muse while my cousin is at work 2011-12-21T15:08:35 "sorry, morning" => "ok, it happens every day" 2011-12-21T15:08:40 thats my best joke ever 2011-12-21T15:08:43 and you dont even get it 2011-12-21T15:08:50 *** thagomizr_ has joined #aichallenge 2011-12-21T15:09:02 Lol, I could repeat 2011-12-21T15:09:04 *** thagomizr has quit IRC (Read error: Connection reset by peer) 2011-12-21T15:09:04 *** thagomizr_ is now known as thagomizr 2011-12-21T15:09:10 Sorry, morning 2011-12-21T15:09:14 :D 2011-12-21T15:09:38 Be Jackson game over? 2011-12-21T15:09:41 its mid-night here :D 2011-12-21T15:09:44 Haha 2011-12-21T15:10:04 It's 7 10am in Sydney 2011-12-21T15:10:20 :P 1:10AM Pakistan 2011-12-21T15:10:27 finally won a game after a big losing streak 2011-12-21T15:10:56 argh, between GreenTea and Memetix, and delineate rolled over for Memetix 2011-12-21T15:11:22 *** Scryer has quit IRC (Remote host closed the connection) 2011-12-21T15:11:28 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-12-21T15:11:35 <_flag> Damn, xathis is pulling away from me 2011-12-21T15:11:36 I won my last two games at least 2011-12-21T15:11:44 Put me back near 40 2011-12-21T15:11:57 *** gaussin has joined #aichallenge 2011-12-21T15:12:15 the stinger: check my profile, I've been doing possibly worse 2011-12-21T15:12:40 you won 2 tho at the top 2011-12-21T15:12:52 *** gustavokambara has quit IRC (Quit: Page closed) 2011-12-21T15:12:54 *** jstemmer has quit IRC (Ping timeout: 255 seconds) 2011-12-21T15:13:05 Yeah I just said that, although I don't think they were much competition 2011-12-21T15:13:16 I had a bunch of games where I placed in the middle 2011-12-21T15:13:20 It's slow for me to look at stuff ATM 2011-12-21T15:13:29 so my mu has been draining away :P 2011-12-21T15:13:31 *** jstemmer has joined #aichallenge 2011-12-21T15:13:40 so deadline extended 1 day? 2011-12-21T15:13:54 *** Ashoka has quit IRC (Quit: Page closed) 2011-12-21T15:13:55 _flag: looking good for you if you can just gold green tea offwe 2011-12-21T15:13:56 http://aichallenge.org/visualizer.php?game=306959&user=2255 weird 2011-12-21T15:14:12 <_flag> Antimatroid: I tested against his bot locally; I think I can do it :) 2011-12-21T15:14:19 this game only showed up on my profile in the past hour or so 2011-12-21T15:14:31 but there were 3 games that appeared before it 2011-12-21T15:14:55 _flag: awesome 2011-12-21T15:15:07 Pesky java programmers 2011-12-21T15:15:16 :p 2011-12-21T15:15:17 whoa i noticed that the cutoff is at 3000 2011-12-21T15:15:20 i lost another game 2011-12-21T15:15:22 ffffffffuck 2011-12-21T15:15:23 where is the next cut 2011-12-21T15:15:32 <_flag> Antimatroid: Haha, they've been doing better and better every contest 2011-12-21T15:15:39 <_flag> In tron there was none of them 2011-12-21T15:15:48 <_flag> Planetwars... we'll they made up some ground 2011-12-21T15:15:56 <_flag> well* 2011-12-21T15:16:01 i think my bot will not make it first (in my country) 2011-12-21T15:16:02 <_flag> Ants if full of them 2011-12-21T15:16:08 *** sashaSochka has joined #aichallenge 2011-12-21T15:17:20 lolx I am only one from my country to reach in top 700's :D 2011-12-21T15:18:15 What country are you from? 2011-12-21T15:18:26 Pakistan 2011-12-21T15:18:41 Australia is slowly getting moreq competitive which is nice 2011-12-21T15:18:52 Smiley is my main competition 2011-12-21T15:18:56 3 poles in top 100 2011-12-21T15:19:11 Aka solifigud 2011-12-21T15:20:16 norway are kind of suck right now, but we've had 2 in top 100 at different times before the final submissions 2011-12-21T15:20:50 Fluxid: youre not gonna be in the top100: 2011-12-21T15:20:53 ? 2011-12-21T15:20:58 Fluxid: but how many zeroes 2011-12-21T15:21:04 jesus h chris 2011-12-21T15:21:17 i can forget about that too then 2011-12-21T15:21:30 *** Antimony has joined #aichallenge 2011-12-21T15:22:03 mcstar: i hope ill stay in top 100 2011-12-21T15:22:14 (i hoped for top 50 before) 2011-12-21T15:22:36 humm, I can't reach the "my profile" page :s 2011-12-21T15:22:46 everything is slow for me now 2011-12-21T15:23:39 what guarantees that TrueSkill has only one fixed-point? 2011-12-21T15:23:45 as in fixed permutation 2011-12-21T15:24:01 Fluxed: I though you were top 30 ish 2011-12-21T15:24:40 Antimatroid: i were before finals ;( 2011-12-21T15:24:54 but suddenly i lose almost every game 2011-12-21T15:25:27 unless a lot of people resubmitted with radically different code I think there's still a lot of upheaval to go in the top 100 2011-12-21T15:25:45 I was around 1500 before finals, and now around 680 :P Last day coding worked perfect :D 2011-12-21T15:26:00 Fluxid: yeah, it got a lot tougher 2011-12-21T15:26:16 I reuploaded a bunch of times, and it got really hard to climb back up 2011-12-21T15:26:28 mcstar : I'm also wondering how it's gonna end. I mean a colleague of mine is better than me on some maps (without water big area) and I'm better on mazes. Is trueskill and mass game supposed to make a "proper" ranking at the end? Or is it going to oscillate again and again? 2011-12-21T15:26:54 yeah, or that 2011-12-21T15:27:06 cause games can differ hugely 2011-12-21T15:27:13 because of different environments 2011-12-21T15:27:26 the skill of a player that the alg. is trying to compute 2011-12-21T15:27:27 Clearly, depending on the number of players, the range at the begining, waters ... 2011-12-21T15:27:30 is not constant 2011-12-21T15:28:03 (by range at the begining, I mean distance between enemy hills) 2011-12-21T15:28:09 Surya: starting positions relative to other bots changes everything 2011-12-21T15:28:17 I still expect a lot of change I. The top 100 before things stablisise to some degree 2011-12-21T15:28:27 Question is, are we goi g to let it stabilize? 2011-12-21T15:28:35 i imagine, in this case, (differring envrionments) the result will be very much dependent upon game-ordering(player matching) 2011-12-21T15:28:37 I think it would beq a shame if we don't 2011-12-21T15:29:04 thestinger: how are you feeling about your rank? 2011-12-21T15:29:08 Antimatroid: its possible that it cant stabilize 2011-12-21T15:29:13 thats what er are getting at 2011-12-21T15:29:16 we* 2011-12-21T15:29:21 BenJackson: I think the 20 to 40 range is where I should be, so it's fine :) 2011-12-21T15:29:23 Yeah, I got the following problem on my bot: I kamikaze a lot when an enemy is close to my hill, so they can't see me nor take my hills. Obviously I don't take in account the number of my ants (sigh) so I lose a lot of games when I'm close from someone 2011-12-21T15:29:31 Are the servers going to shut down at the 'winner deadline'? 2011-12-21T15:29:33 BenJackson: I had some unlucky games, but now I've had 2 lucky ones 2011-12-21T15:29:36 Mcstar: sure, but I think it can a lot more than currently 2011-12-21T15:29:47 BenJackson: unhappy with the maps tho :P 2011-12-21T15:29:52 yeah 2011-12-21T15:29:57 the maps could be better 2011-12-21T15:30:06 on the one hand I like new random maps 2011-12-21T15:30:07 A lot better :p 2011-12-21T15:30:21 on the other selecting "good"maps would have been nice 2011-12-21T15:30:26 only played on one really big open map, which my bot does well on 2011-12-21T15:30:45 thestinger: the phase I'm in now (coming down from 1st) is the part where my bot is whipped back to the right rank 2011-12-21T15:30:52 I knew it was coming but it's still hard to watch :) 2011-12-21T15:31:09 well, we both got a chance to hold #1 2011-12-21T15:31:35 I got to hang onto it for almost a whole day due to the slow game rate for the top bots :) 2011-12-21T15:31:49 I don't like hand selecting maps, it adds human bias as to which aspects we want to give an advantage 2011-12-21T15:31:49 *** sashaSochka has quit IRC (Quit: Page closed) 2011-12-21T15:32:07 That is unavoidable but I'd rather capture that as best as possible in the generator 2011-12-21T15:32:29 thestinger: apparently I was #1 again briefly after xathis took it 2011-12-21T15:32:31 which surprised me 2011-12-21T15:32:38 didn't get that screenshot :) 2011-12-21T15:32:43 xathis is pulling way ahead now 2011-12-21T15:32:48 BenJackson: there are irc logs to prove it :) 2011-12-21T15:34:42 *** taeke has quit IRC (Quit: Page closed) 2011-12-21T15:35:24 BenJackson, now that you can see the design of my bot, what do you think of it? 2011-12-21T15:36:19 hah, I haven't read ANY of them yet 2011-12-21T15:36:22 68.79->69.46 2011-12-21T15:36:23 what's your url? 2011-12-21T15:36:28 i just love my red notification 2011-12-21T15:36:51 thestinger: do you have a notification? 2011-12-21T15:36:55 sending texts to my phone now when game count changes (with mu, sigma and some other stuff) 2011-12-21T15:37:05 my script has printed out (since I started watching) 2, 4, 4, 4, 8, 5, 6, 5, 6, 9, 10 2011-12-21T15:37:11 everyone sends messages to their phone 2011-12-21T15:37:13 wth? 2011-12-21T15:37:18 well, I started with emails 2011-12-21T15:37:28 arent yoiu before your computer? 2011-12-21T15:37:31 but then I just switched to the email that goes to my phone as a text message 2011-12-21T15:37:35 before->in front of 2011-12-21T15:38:20 not always :P 2011-12-21T15:38:26 mcstar: I've done 50% of MyBot.ml 2011-12-21T15:38:35 great 2011-12-21T15:38:42 im going to get that printer 2011-12-21T15:38:46 *** grwip has quit IRC (Ping timeout: 248 seconds) 2011-12-21T15:39:08 Here's the thread for my bot. http://forums.aichallenge.org/viewtopic.php?f=24&t=2169 2011-12-21T15:39:53 *** dici has quit IRC (Quit: ( www.nnscript.com :: NoNameScript 4.22 :: www.esnation.com )) 2011-12-21T15:40:05 anyway, there's a whole API designed just for me :P 2011-12-21T15:40:09 http://aichallenge.org/ranking_json.php?language_id=25&country_id=2&org_id=346 2011-12-21T15:41:00 *** epicmonkey has joined #aichallenge 2011-12-21T15:41:08 *** cyphase has quit IRC (Read error: Connection reset by peer) 2011-12-21T15:41:16 oh? 2011-12-21T15:41:28 well no, but it returns 1 result for the query to get my info 2011-12-21T15:41:30 Antimony: I used simulated annealing which is similar to a constraint solver 2011-12-21T15:41:31 why wasnt ranking_json.php made public? 2011-12-21T15:41:36 *** xyxan has quit IRC (Ping timeout: 258 seconds) 2011-12-21T15:41:36 it just maximizes a goal function 2011-12-21T15:41:38 for BenJackson it returns like 33KB of crap xD 2011-12-21T15:41:45 *** xyxan has joined #aichallenge 2011-12-21T15:41:59 Antimony: and I stole the idea for possible ants and it worked quite well (I think I heard it from you on IRC) 2011-12-21T15:42:12 and I also use influence functions all over, mostly derived from global bfs's 2011-12-21T15:42:21 mcstar: https://github.com/aichallenge/aichallenge/blob/epsilon/website/ranking_json.php 2011-12-21T15:42:39 if I had it to do again I'd compute a lot of the same influence info but use a global solver for territory 2011-12-21T15:42:40 I just got the ids by fetching the first page and searching for my entry 2011-12-21T15:42:52 (my ant spread-out is brute force in the annealing score function) 2011-12-21T15:42:56 I think I should have added some randomness into my constraint solver, to make it more like annealing 2011-12-21T15:43:11 greedy search is good for finding a solution, but terrible for exhaustive search 2011-12-21T15:43:15 I wrote a maximum matching implementation (the O(n^3) "fast" version) 2011-12-21T15:43:19 but I didn't actually use it 2011-12-21T15:43:31 thestinger: so i can get the same with manual search for the right org_id? 2011-12-21T15:43:56 *** meduza has joined #aichallenge 2011-12-21T15:43:57 BenJackson: i use it all over the place :D 2011-12-21T15:43:59 mcstar: yeah, figure out which page you're on in the rankings, then do ranking_json.php?page=11 or w/e 2011-12-21T15:44:07 and find your org_id, country_id and language_id 2011-12-21T15:44:12 yeah 2011-12-21T15:44:48 C++11 narrows it down a lot 2011-12-21T15:45:28 whenever i want to move a bunch of ants, i use hungarian matching 2011-12-21T15:45:35 for example, battle 2011-12-21T15:45:46 the rows are the ants 2011-12-21T15:45:52 the cols are the possible next tiles 2011-12-21T15:46:05 so it guarantees that no two ant will step an 1 tile 2011-12-21T15:46:09 on 2011-12-21T15:46:32 I implemented Kuhn Munkres 2011-12-21T15:46:51 which is the graphy-version of hungarian that's O(n^3) instead of O(n^4) 2011-12-21T15:47:06 mcstar: +1 for hungarian 2011-12-21T15:47:21 i use it for expansion too 2011-12-21T15:47:42 i didnt write mine 2011-12-21T15:47:46 i found it on the net 2011-12-21T15:47:52 the author is in the source 2011-12-21T15:48:14 it was quite fast so i didnt bother with making mine, but i probably should have 2011-12-21T15:48:20 for experience 2011-12-21T15:48:25 mcstar: we have also used hungarian, except that we used it for fighting: it gave the initial direction for ants to move, which we passed to genetic search 2011-12-21T15:48:49 BenJackson: kuhn munkers and hungarian method are the same thing. aren't the? 2011-12-21T15:49:01 no 2011-12-21T15:49:09 like he said kuhn is faster 2011-12-21T15:49:23 the very first alg. was called hungarian 2011-12-21T15:49:31 and those guys imporved on it 2011-12-21T15:49:33 meduza: hungarian is the row elimination one 2011-12-21T15:49:58 BenJackson: oh, then i also use kuhn munkers 2011-12-21T15:49:58 kuhn munkres uses augmenting paths from an alternating tree 2011-12-21T15:50:10 https://github.com/liquid-phynix/aicbot/blob/master/hungarian.cpp 2011-12-21T15:50:28 my distinction may be incorrect, but it was the only one I could find that turned up the right literature 2011-12-21T15:50:28 Damn, I didn't know about that :( 2011-12-21T15:50:37 BenJackson: yep, i thought it was initially introduced in hungarian algorithm 2011-12-21T15:50:39 Antimatroid: ? 2011-12-21T15:50:50 Antimatroid: i was telling you about this in april 2011-12-21T15:50:54 Hungarian algo 2011-12-21T15:50:56 I ended up buying _Graphs, Networks and Algorithms_ by Dieter Jungnickel because it had such a good description of the algorithm 2011-12-21T15:50:58 Lol 2011-12-21T15:51:02 Always 2011-12-21T15:51:10 remember? i did similar thing in common lisp 2011-12-21T15:51:17 and i had an implementation there 2011-12-21T15:51:27 Nope :p 2011-12-21T15:51:28 we used this implementation - one of the best - http://acm.mipt.ru/twiki/bin/view/Algorithms/HungarianAlgorithmCPP 2011-12-21T15:51:30 i used it for optimal food gathering 2011-12-21T15:51:34 *** teapotahedron_ni has quit IRC (Ping timeout: 258 seconds) 2011-12-21T15:51:41 i assigned ants to foods 2011-12-21T15:51:44 the code is on the bottom of the page 2011-12-21T15:51:48 the cost was the manhattan distance 2011-12-21T15:51:54 *** itzkoow has joined #aichallenge 2011-12-21T15:51:56 Can we start the contest over please ? 2011-12-21T15:52:03 :D 2011-12-21T15:52:11 damn, i know i said it to you 2011-12-21T15:52:18 i cant be blamed that you dont remember 2011-12-21T15:52:22 You probably did 2011-12-21T15:52:33 I'm terrible at these kind of things 2011-12-21T15:53:19 I was semi keen on getting one of these iPads, till I found out you can't install compilers 2011-12-21T15:53:26 BenJackson: you bought that book for the contest? 2011-12-21T15:53:31 i just hope next challenge will be as good as this one 2011-12-21T15:53:42 Antimatroid: hack it to use linux 2011-12-21T15:53:58 You can do that? 2011-12-21T15:54:02 I might do that then 2011-12-21T15:54:05 google around 2011-12-21T15:54:10 *** kilae has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 8.0/20111104165243]) 2011-12-21T15:54:11 it might be possible 2011-12-21T15:54:15 mcstar: are you a lisper? 2011-12-21T15:54:19 or just get another tablet that supports it 2011-12-21T15:54:28 Hey mega1 :p 2011-12-21T15:54:37 helloo 2011-12-21T15:54:38 mega1: hi, i started learning it because i saw your winning 2011-12-21T15:54:39 *** teapotahedron_ni has joined #aichallenge 2011-12-21T15:54:41 mcstar: I bought it because I found the page for that algo on google books (?) and it was a much better presentation than most graph algorithm books 2011-12-21T15:54:51 Where have you been this contest? 2011-12-21T15:55:03 mega1: kezdo, de remelem makd egyszer kozephalado :) 2011-12-21T15:55:13 *** itzkoow has quit IRC (Client Quit) 2011-12-21T15:55:16 mcstar: ooh 2011-12-21T15:55:21 do I know you? 2011-12-21T15:55:31 no i dont think so 2011-12-21T15:55:47 please explain like i'm 5: hungarian algorithm 2011-12-21T15:55:53 *** amstan_ is now known as amstan 2011-12-21T15:55:55 Antimatroid: keeping away :-) 2011-12-21T15:55:57 mcstar: this one helped me a lot http://www.nada.kth.se/~johanh/ (last item in links section) - really good notes 2011-12-21T15:56:05 Mega1 wise choice 2011-12-21T15:56:13 *** ltriant has joined #aichallenge 2011-12-21T15:56:16 *** Antimony has quit IRC (Ping timeout: 240 seconds) 2011-12-21T15:56:36 I hope I have more time for next contest 2011-12-21T15:57:05 meduza: ill try to write my own, but i didnt had enough trouble with my bot 2011-12-21T15:57:25 where i say didnt i actually mean did and vica versa 2011-12-21T15:57:27 *** cyphase has joined #aichallenge 2011-12-21T15:59:19 mega1: a remark: next time please announce your winning in hungarian also, and for a bigger audience 2011-12-21T15:59:29 mcstar: wikipedia article of the "hungarian algorithm" has a general description in graph theory terms. i implemented it based on that definition 2011-12-21T15:59:56 what about the domination of xathis? Do you know how did he achieve that? 2011-12-21T16:00:12 mega1: im just disappointed seeing so few fellow countrymen after PW 2011-12-21T16:00:15 mcstar: next time :-) sounds good 2011-12-21T16:00:30 mega1: nothing really special, lots of bfs 2011-12-21T16:00:58 xathis: post mortem in the works, I take it? 2011-12-21T16:01:22 mega1: yes, i try to have it ready when the finals end 2011-12-21T16:02:36 *** McLeopold has joined #aichallenge 2011-12-21T16:03:05 *** Kurnevsky has quit IRC (Quit: Instantbird 1.0) 2011-12-21T16:06:22 *** olexs has joined #aichallenge 2011-12-21T16:07:20 I beat cumbuz 2011-12-21T16:07:22 good, do it while the adrenalin lasts 2011-12-21T16:07:39 But they don't currently have a very good ranking 2011-12-21T16:08:45 *** QGazQ has quit IRC (Read error: Connection reset by peer) 2011-12-21T16:09:01 *** QGazQ has joined #aichallenge 2011-12-21T16:09:11 oh shit, big loss 2011-12-21T16:09:12 *** itzkow has quit IRC (Ping timeout: 258 seconds) 2011-12-21T16:09:27 sadly I am the stepping stone by which chrish and pguillory get sorted 2011-12-21T16:09:57 wow that game was quite close 2011-12-21T16:10:10 BenJackson ouch 2011-12-21T16:10:38 *** Manwe56 has joined #aichallenge 2011-12-21T16:10:50 good news is that at 16th I'll probably get some easy games next 2011-12-21T16:11:17 BenJackson, not luck.. 2011-12-21T16:11:50 GreenTea: ? 2011-12-21T16:12:26 *** Antimony has joined #aichallenge 2011-12-21T16:12:31 *** dmj111 has quit IRC (Ping timeout: 276 seconds) 2011-12-21T16:12:43 BenJackson, you have big growing count of ants in that game 2011-12-21T16:13:21 oh, yes, in mazes sometimes there's just not enough time 2011-12-21T16:14:29 mcstar: I've done with MyBot.ml. It even compiles... 2011-12-21T16:14:52 what do you guys think about this? http://forums.aichallenge.org/viewtopic.php?f=27&t=2198 2011-12-21T16:16:08 thestinger: http://aichallenge.org/visualizer.php?game=308886&user=53 heheh, this is backdoor! Not fair :D 2011-12-21T16:16:30 amstan: C#? GC? 2011-12-21T16:16:48 maybe the engine rules should have specifically allowed for gc 2011-12-21T16:16:58 Ams tan could be a serious problem 2011-12-21T16:16:59 he wants me to let him change his settings so he doesn't timeout anymore 2011-12-21T16:17:00 *** Antimony has quit IRC (Ping timeout: 244 seconds) 2011-12-21T16:17:10 My bot is c# and I had no timeout because I test it all the time. But it's not a proof his code is bad (my 2 cents ;)) 2011-12-21T16:17:10 Probably worth at least investigating 2011-12-21T16:17:21 eg if a bot says "gc" instead of "go" it gets an extra turn between turns where it calls System.gc() 2011-12-21T16:17:22 No change of settings 2011-12-21T16:17:41 there hasn't been any changes on the workers since dec 10th 2011-12-21T16:17:42 amstan: if you can find some systematic problem I'd say fix it 2011-12-21T16:17:47 although getting him scored right is a big problem 2011-12-21T16:17:58 mcstar: http://ompldr.org/vYnVzaw/MyBot.ml 2011-12-21T16:18:02 if he wants to change code due to timeouts, why did he not find this before finals? 2011-12-21T16:18:14 sry, cannot process ants.ml today 2011-12-21T16:18:23 BenJackson: please tell him that 2011-12-21T16:18:24 Has he let someone else run the code to see if it actually does/n't timeout? 2011-12-21T16:18:41 Antimatroid: so i'm not saying that the timeouts are not real 2011-12-21T16:19:04 but there's nothing that can be done, everyone's equally(same language) affected by them 2011-12-21T16:19:06 Is there a way to know for sure it's not his code causing problems? Maybe new maps make bugs suddently appear? 2011-12-21T16:19:17 he has submits on 14, 16, 17 17 18 19 2011-12-21T16:19:20 gc :O :O 2011-12-21T16:19:39 *** Hexren has joined #aichallenge 2011-12-21T16:19:42 *** Hexren has left #aichallenge 2011-12-21T16:19:54 I see many timeouts for v17 2011-12-21T16:20:09 *** McLeopold has quit IRC (Ping timeout: 240 seconds) 2011-12-21T16:20:09 *** dr- has quit IRC (Ping timeout: 240 seconds) 2011-12-21T16:20:10 Surya: you can profile with the input data from the bot 2011-12-21T16:20:30 amstan: I get frustrated with people who say "my code works at home but not on the server! the servers are broken!" 2011-12-21T16:20:54 *** jab_bott has joined #aichallenge 2011-12-21T16:20:55 Amstan: yeah I just mean if something on the server side is going wrong for them to time out then we have a problem, if it's just an issue with their bot then that's their problem and I don't think we allow them to change anything 2011-12-21T16:21:02 my personal policy is that if you pick a language that has unpredictable GC, you have to accept all the problems that it might happen because of it 2011-12-21T16:21:11 amstan: my bot experiencing timeout issues sometimes (twice for the finals), but i'm not sure that it's not an implementation bug :) 2011-12-21T16:21:12 tbh, I do think the servers aren't really stable. 2011-12-21T16:21:17 amstan: or not accept, perhaps, but work around 2011-12-21T16:21:22 ChrisH had to with Go 2011-12-21T16:21:31 *** chris___0076 has joined #aichallenge 2011-12-21T16:21:32 mleise: did you check https://github.com/aichallenge/aichallenge/pull/390 ? (maybe I should put it in a separate pull request) 2011-12-21T16:21:33 i also see timeouts for lama3oid on Dec 13, 14, 15, 17 2011-12-21T16:21:43 amstan: but likely it's not a bug 2011-12-21T16:21:44 a bunch of them 2011-12-21T16:21:44 yes, like reuse the same data structures and don't rely on gc 2011-12-21T16:22:07 Yeah, I mean if there was literally an issue with the server we don't know about, but that seems unlikely to me, there aren't that many people complaining 2011-12-21T16:22:20 I put the gc off in python :v 2011-12-21T16:22:29 one of the main reasons I went straight from python to c++ was that I wanted to be 110% sure that I could manipulate grids of things without any risk of generating tons of garbage 2011-12-21T16:22:33 amstan: we should throw away our java and .net and code in c 2011-12-21T16:22:47 will there be any special battles - for example reduce the turntime and see who works best at very low processing requirements.. and play with visible/attack radii to see whose bots are the most adaptable? 2011-12-21T16:22:54 <_flag> If xathis doesn't timeout no other java bot should timeout, right? 2011-12-21T16:23:02 Riiiight. 2011-12-21T16:23:04 @rankings 2011-12-21T16:23:05 Antimatroid: Top 10 players: xathis(88.8), teapotahedron(84.9), FlagCapper(84.9), meduza(82.8), runevision(82.7), Komaki(82.3), Memetix(82.2), delineate(81.4), GreenTea(81.2), Speedy_Consoles(81.1) 2011-12-21T16:23:09 <_flag> Because clearly there is a way to do things better without timeing out 2011-12-21T16:23:15 Look out flag 2011-12-21T16:23:31 <_flag> Antimatroid: This is the game I just "lost" http://aichallenge.org/visualizer.php?game=309495&user=31 2011-12-21T16:23:37 I had quite bad problems with gc, even doing system.gc etc didn't help. Ran it through profiler something to do with allocating (and so needing GC) on around 100MB per turn 2011-12-21T16:23:52 Changed the code so I had a few pools of objects instead and the problem went away 2011-12-21T16:24:17 As I say I might not have a good bot, but it has taught me things about java (day job I'm c++) 2011-12-21T16:25:16 _flag: my bot is good at dominating with ants yet placing way down 2011-12-21T16:25:31 *** chris__0076 has quit IRC (Ping timeout: 276 seconds) 2011-12-21T16:26:01 *** Antimatroid has quit IRC (Quit: Page closed) 2011-12-21T16:27:52 am surprised a1k0n is at 19th 2011-12-21T16:27:55 bretep: cpython does reference counting anyway 2011-12-21T16:28:02 the GC just takes care of cyclical references 2011-12-21T16:28:18 lots of my versions used up almost all the turntime 2011-12-21T16:28:45 moving each 100 ants takes a minimum of 10ms due to how the engine reads it in 2011-12-21T16:28:52 *** Antimatroid has joined #aichallenge 2011-12-21T16:28:55 so I gave it 20ms per 100 ants 2011-12-21T16:29:25 my final version doesn't use a timer tho because using up all the time to generate and evaluate more enemy moves didn't help 2011-12-21T16:29:48 Yer in java I had to rewrite the input parcer a bit, the one in the starter bot took almost all the time on multi hill maps on the first turn 2011-12-21T16:29:52 thestinger: what rank are you at then? 2011-12-21T16:29:59 UncleVasya: ok thx, i was shaving 2011-12-21T16:30:01 It would be interesting to analyze the # of players ants at timeout to see how many people got that wrong 2011-12-21T16:30:03 amstan: now? :) 2011-12-21T16:30:09 http://aichallenge.org/profile.php?user=2255 2011-12-21T16:30:19 thestinger: well.. more accurate than that 2011-12-21T16:30:27 thestinger: not badf 2011-12-21T16:30:33 Also when do most people give up time wise? To be on the safe side I stop and do the output and a few tidy up bits when theres only 100ms left 2011-12-21T16:30:34 I was delighted to get up to 96th - but quickly went out of the top 100 again, back down at 180ish again now I think 2011-12-21T16:30:34 for a bot without timers.. 2011-12-21T16:30:45 it had one :P 2011-12-21T16:30:46 mcstar, antimatroid, I've lost only 1 ant (and in 1-to-1 excahnge) but strcat stiil got a victory: http://aichallenge.org/visualizer.php?game=308886&user=53 2011-12-21T16:30:51 <_flag> jab_bott: You wrote the visualizer overlay, right? 2011-12-21T16:30:53 Love such games 2011-12-21T16:30:54 *** Varan has joined #aichallenge 2011-12-21T16:31:07 amstan: I just adjust the combat score so I sacrifice a higher ratio of ants if I have lots 2011-12-21T16:31:12 _flag: yeah.. well I added some bits to the existing visualizer 2011-12-21T16:31:14 thestinger: you might be the top bot which doesn't have a timer 2011-12-21T16:31:37 BenJackson: did you end up needing a timer? 2011-12-21T16:31:40 <_flag> jab_bott: Right, just wanted to let you know, I've been trying to use it and the java version has a tendency to crash after setting a lot of tiles 2011-12-21T16:31:48 thestinger: no 2011-12-21T16:31:52 <_flag> Not sure if you wanted to look at it or not 2011-12-21T16:31:53 amstan: ^^ 2011-12-21T16:32:03 BenJackson new highest bot who didn't need a timer! 2011-12-21T16:32:12 hehe 2011-12-21T16:32:12 well let's not say *need* 2011-12-21T16:32:17 until I avoid all timeouts :) 2011-12-21T16:32:29 _flag: yeah - I never really used the streaming version... the whole thing was a hack anyway :D Hopefully next time we'll have visualization built into the engine right from the start :) 2011-12-21T16:32:34 amstan: I had it in the code until the end, but it was pointless because I wasn't using all the time 2011-12-21T16:32:43 use* 2011-12-21T16:32:55 I didn't organize my code in a way that I could stop early 2011-12-21T16:33:03 and since I wasn't timing out I didn't bother to restructure 2011-12-21T16:33:03 heheh 2011-12-21T16:33:10 <_flag> jab_bott: Do you know of any way I can output a game to images in order to create a video like it was done with the java version? 2011-12-21T16:33:11 I think I timed out in one game during the run of play 2011-12-21T16:33:23 and that turn was not possibly 500ms so it was the server 2011-12-21T16:33:32 earlier turns in the same game took 5x longer when I re-ran the input 2011-12-21T16:33:35 *** raja_ has joined #aichallenge 2011-12-21T16:33:36 http://aichallenge.org/profile_games.php?user=453&page=3 just sort it by status. It's more then enough to understand that problem is not in servers. 2011-12-21T16:33:44 _flag: sure you can just write data to an image during your turn and output that 2011-12-21T16:33:58 whoa someone wiped the tc pants 2011-12-21T16:34:12 <_flag> jab_bott: Yes, but I'd have to reimplement a lot of code, I meant using the existing visualizer 2011-12-21T16:34:26 amstan: oh, I had a version with an infinite loop bug - so there's a timeout with that one 2011-12-21T16:34:32 lol 2011-12-21T16:34:43 I did that a few times in development 2011-12-21T16:34:43 *** u_ has joined #aichallenge 2011-12-21T16:34:45 _flag: oh.. maybe using the replay data you could yeah - the visualizer writes lots of stuff to images anyway 2011-12-21T16:34:50 it would adjust my offensive move and then readjust to the old move 2011-12-21T16:34:56 I also had a few crashes in development due to obscure cases like unreachable hills 2011-12-21T16:35:00 _flag: you could just record it.. 2011-12-21T16:35:06 oh or having 0 hills 2011-12-21T16:35:07 _flag: why do you need this anyway? 2011-12-21T16:35:25 <_flag> amstan: I'd like to have a visual component to my writeup-type-thing 2011-12-21T16:35:30 <_flag> So I can easily show what the bot is doing 2011-12-21T16:35:33 so just include a visualizer 2011-12-21T16:35:37 *** iglo has quit IRC (Remote host closed the connection) 2011-12-21T16:35:50 _flag: yeah, screen-capture a playback... or if you really want exact images you could add code to the canvas to write the images to file 2011-12-21T16:36:06 *** tomb___ has quit IRC (Quit: Page closed) 2011-12-21T16:36:08 _flag: you could have anchors which make the visualizer to to a specific turn 2011-12-21T16:36:14 it's fairly scriptable i think 2011-12-21T16:36:17 <_flag> amstan: I'm adding enough different things to the visualization that it becomes fairly slow and it would be nice to capture it 2011-12-21T16:36:44 _flag: can't you modify the js to emit pngs from the canvas itself? 2011-12-21T16:36:50 other than the jar i'm not sure 2011-12-21T16:36:55 contestbot: seen mleise 2011-12-21T16:36:55 amstan: mleise was last seen in #aichallenge 1 day, 2 hours, 26 minutes, and 16 seconds ago: but Migi32, that was your idea for this contest already ;) 2011-12-21T16:36:56 _flag: yeah speed is a massive problem, the communication mechanism is really slow and uses up lots of data 2011-12-21T16:36:59 <_flag> BenJackson, jab_bott: That's the next thing I'm going to try, thanks 2011-12-21T16:37:02 ... 2011-12-21T16:37:15 okay, when i get up in the morning 2011-12-21T16:37:22 mleise: hey, _flag wants to make the jar output videos, but he can't get jab_bott's debug stuff going in it 2011-12-21T16:37:25 I WANT TO BE IN TOP 50 OKAY? 2011-12-21T16:37:27 ;) 2011-12-21T16:37:37 <_flag> mleise: Do you know of an easy place I can insert code into the javascript to output each frame to a png/jpeg? 2011-12-21T16:37:39 Fluxid: how about we just call you the winner? 2011-12-21T16:37:45 *** grwip has joined #aichallenge 2011-12-21T16:38:01 Ok with that, if I can be second 2011-12-21T16:38:06 amstan: that would be too good 2011-12-21T16:38:07 _flag: no, not really. the video output is mostly done on the Java side 2011-12-21T16:38:18 g'nite 2011-12-21T16:38:37 mleise: _flag: surely CanvasElement.js could write the images out? I think it composites everything into an image anyway 2011-12-21T16:39:16 <_flag> Can javascript even create files? 2011-12-21T16:39:19 a "save as movie" with a turn range would be awesome 2011-12-21T16:39:33 _flag: if you load as file:// maybe? 2011-12-21T16:39:39 jab_bott: JavaScript cannot write images to disk, but on the Java side, you are correct, I write the composited image to disk using the JRE. 2011-12-21T16:39:54 ahhhh right 2011-12-21T16:40:04 *** mj41 has joined #aichallenge 2011-12-21T16:40:37 <_flag> mleise: So does the canvas visualizer use java code or just js? 2011-12-21T16:40:39 You'll have to take a look at the video option and see if it meets your expectations. 2011-12-21T16:40:46 amstan: so how about some special side-contests to see which bots can cope with parameters changing like different attack radius and shorter turn times? 2011-12-21T16:40:48 *** kire has quit IRC (Remote host closed the connection) 2011-12-21T16:40:53 Crap, I think my bot might get semi crippled when it has no hills left 2011-12-21T16:40:55 The canvas uses JS *only* 2011-12-21T16:41:06 <_flag> mleise: Because the java visualizer with jab_bott's additions crashes on me so I can't even use the video output feature 2011-12-21T16:41:08 jab_bott: go right ahead, get the tcp server code and modify it 2011-12-21T16:41:11 <_flag> Only the canvas one works 2011-12-21T16:41:14 canvas uses js, there is also a jar visualizer 2011-12-21T16:41:16 mleise: the problem is that with lots of visualizations, the java one bites the dust 2011-12-21T16:41:38 jab_bott: I am aware of that. 2011-12-21T16:41:47 jab_bott: try giving it more memory 2011-12-21T16:41:51 jab_bott: or try sun java 2011-12-21T16:42:09 The Java visualizer wraps the JS visualizer, simulating a web environment (as far as I needed it) using Rhino (from Mozilla) 2011-12-21T16:42:15 hm, nothing weird seems to be happening near rank 3000 2011-12-21T16:42:28 * amstan checks rank 3000 just in case 2011-12-21T16:42:56 JavaScript inside Java is a major pita: it is slow and uses insane amounts of memory 2011-12-21T16:43:16 nope.. just a lot of bots with the same skill 2011-12-21T16:43:26 I couldn't forsee the number of ants though 2011-12-21T16:43:28 amstan: I was looking for a gap or some other oddness 2011-12-21T16:43:40 yeah.. just give java a lot of memory, give it some swap 2011-12-21T16:43:50 _flag: what does the Java visualizer crash on? 2011-12-21T16:43:51 and let it do its thing for a few hours 2011-12-21T16:43:55 amstan: pos 3000 lost SIX games with _Nord_ 2011-12-21T16:43:55 *** ChrisH_ has joined #aichallenge 2011-12-21T16:44:08 BenJackson: who's that? 2011-12-21T16:44:19 http://aichallenge.org/profile.php?user=5043 2011-12-21T16:44:31 who's nord? 2011-12-21T16:44:32 mleise: for the next contest (when is it anyway?) maybe we can build bot-supplied visualizations right into the engine? perhaps provide a library that bots can use to write images/video straight out during play too? or at the least have a fast binary interface for sending lots of data to be rendered 2011-12-21T16:44:35 oh they're tying for 1st 2011-12-21T16:44:38 amstan: nobody 2011-12-21T16:44:42 weird.. 2011-12-21T16:44:45 just seemed odd one guy would play 6 games with him 2011-12-21T16:44:56 <_flag> mleise: It crahses as soon as I activate jab_bott's visualizer overlay addition 2011-12-21T16:44:59 where's mcleopold with a "who's played a lot with who" query 2011-12-21T16:45:09 can i borrow an usb printer cable? 2011-12-21T16:45:24 BenJackson: sorry about that last game, but I'm glad for the bump in my mu. 2011-12-21T16:45:25 mcstar: sure.. 2011-12-21T16:45:32 *** ert has quit IRC (Quit: Page closed) 2011-12-21T16:45:33 * bretep throws a usb cable to poland 2011-12-21T16:45:43 Got it, mcstar ? 2011-12-21T16:45:56 ChrisH_: I came in last (or more like late!) but it was a good game 2011-12-21T16:46:12 I don't mind losing as much as losing stupidly 2011-12-21T16:46:37 _flag: ah i think it uses some kind of code injection 2011-12-21T16:46:43 yeah, pretty much a standoff, and my spawn control was the difference that let me keep my isolated hill alive until the turn limit 2011-12-21T16:47:00 that will never work 2011-12-21T16:47:38 All the JS is compiled to Java byte code to make it run faster, so no dynamic loading of JS code. 2011-12-21T16:47:46 jab_bott, mleise, _flag: so you are all here, just put it together and you might be able to fix it 2011-12-21T16:47:48 BenJackson: my bot was starting to fade at the end though, so i wouldn't have won if it had gone longer 2011-12-21T16:47:53 *** Jak_o_Shadows has joined #aichallenge 2011-12-21T16:48:22 _flag: I've had replay files over 150M due to visualizations - are you familiar with how it works? 2011-12-21T16:48:36 <_flag> mleise: It works with simple additions, and crashes after a certain type of addition, so it might be due to his code 2011-12-21T16:48:53 so some custom visualizations work? 2011-12-21T16:48:59 then what is the error message? 2011-12-21T16:49:00 <_flag> Some of them work, yes 2011-12-21T16:49:06 ChrisH_: originally V7 was very good early (won lots of multi-hill mazes) and later I optimized for surviving/winning long games 2011-12-21T16:49:07 _flag: basically it sends all the commands you output into the JSON and the visualizer parses them.. not very elegant or fast, but quick to add to the existing visualizer 2011-12-21T16:49:16 I never really re-checked my early game performance 2011-12-21T16:49:25 *** mega1 has quit IRC (Ping timeout: 255 seconds) 2011-12-21T16:49:35 jab_bott: so there is no 'eval' in the code? 2011-12-21T16:49:35 partly because xathis got away with it :) 2011-12-21T16:49:50 mleise: no - there was originally for the first iteration 2011-12-21T16:49:55 All you new guys know who mega1 is right? 2011-12-21T16:50:01 <_flag> It complains that the alpha colour parameter is out of range, but it's valid on the canvas one 2011-12-21T16:50:04 BenJackson: my bot seems to be doing better on maps with fewer players in finals. It is struggling on the 8,9 and 10 player maps. 2011-12-21T16:50:09 :P on top of profile page "until winner announced" OR "untill winner will be anounced" :P janzert amstan contestbot 2011-12-21T16:50:09 Aka bocsimaco 2011-12-21T16:50:12 mleise: then I refactored it so that only the strings are sent and the JS processes them 2011-12-21T16:50:25 ChrisH_: too bad there are no points for holding out the longest in those huge multi-hill chaos maps 2011-12-21T16:50:30 jab_bott: ok, that was where i was wrong 2011-12-21T16:50:31 mleise: I'd never used JS or python before so I was making it up as I went along :P 2011-12-21T16:50:32 I played one where both my neighbors dove on me 2011-12-21T16:50:34 I held out 2011-12-21T16:50:39 but meanwhile someone else expanded 2011-12-21T16:50:51 :P on top of profile page "until winner announced" OR "untill winner will be anounced" :P 2011-12-21T16:50:52 yeah, that has happened to me a few times too 2011-12-21T16:51:20 _flag: maybe the cause is some JS HTML5 canvas functions that are not implemented in the fake web environment? mleise does this sound plausible? 2011-12-21T16:51:25 if there really are a preponderance of 2-player random walks I look forward to that phase of the contest :) 2011-12-21T16:51:32 Antimatroid: and he is not participating, right? 2011-12-21T16:51:45 Didn't participate 2011-12-21T16:52:05 He tried to stay away from planet wars then won it :p 2011-12-21T16:52:54 jab_bott: yes that sonds plausible. I hit that problem a few times myself 2011-12-21T16:52:56 BenJackson: yeah, I won two of those earlier, so maybe that will be my salvation as well. 2011-12-21T16:53:07 _flag: if that's the case, maybe adding clamps to your alpha values to ensure they're from 0.0 to 1.0 might help? 2011-12-21T16:53:10 Antimatroid: could not resist :) 2011-12-21T16:53:33 without the error message we can't tell which function is not implemented, though :) 2011-12-21T16:54:09 *** jstemmer has quit IRC (Ping timeout: 255 seconds) 2011-12-21T16:54:10 mleise: could be arc (used to draw circle) 2011-12-21T16:54:55 *** jstemmer has joined #aichallenge 2011-12-21T16:56:39 amstan: slide cutoff again soon? 2011-12-21T16:56:50 BenJackson: not sure, janzert's department 2011-12-21T16:56:57 janzert: ^^ 2011-12-21T16:57:02 <_flag> mleise, jab_bott: Actually, now I'm getting a different error. "org.mozilla.javascript.EcmaError: TypeError: Cannot call method "toLowerCase" of undefined (CanvasElement.js#949)" 2011-12-21T16:57:16 janzert if so to what value 2011-12-21T16:57:46 ChrisH_: I was telling my wife about that last game and I realized that it probably was bots 2,3,4,5 or close all in one game finishing roughly in the correct order 2011-12-21T16:57:46 *** Antimony has joined #aichallenge 2011-12-21T16:57:59 _flag: sorry the vis add ons were not a whole load better - I came pretty late to the contest and spent half my time on the visualizer, but then I gave up to spend some time on my bot once it got good enough... it sure helped me to find problems in my bot though so was worth the time and glad I could share it so easily and glad otheres used it :) 2011-12-21T16:58:03 *** UncleVasya has quit IRC (Ping timeout: 252 seconds) 2011-12-21T16:58:09 *** mj41 has quit IRC (Ping timeout: 268 seconds) 2011-12-21T16:58:32 <_flag> jab_bott: Not a problem, if it wasn't there I'd have to write it myself, so :) 2011-12-21T16:58:53 *** Hexren has joined #aichallenge 2011-12-21T17:00:07 *** Hexren has quit IRC (Client Quit) 2011-12-21T17:00:08 BenJackson: I just noticed a couple of interesting things about that last game. 1. My bot actually connect the territory between all of my hills by later in the game. 2. I think my bot was struggling with the time constraint. I saw some of my fighting ants stop moving and then get killed by lazarant right my peak of ant population. 2011-12-21T17:00:41 *** Lithosphere has quit IRC (Remote host closed the connection) 2011-12-21T17:00:58 BenJackson: oh, sorry, we just played again 2011-12-21T17:00:59 *** Lithosphere has joined #aichallenge 2011-12-21T17:02:11 BenJackson: I noticed some weird times before the finals started where my bot played the same game twice in a row, as in picked the same opponents one minute after 2011-12-21T17:02:48 ChrisH_: probably a good test of whose combat is more efficient 2011-12-21T17:02:50 I know that's not mine 2011-12-21T17:03:01 jab_bott: yeah, I posted to the forum about that once 2011-12-21T17:03:47 *** amstan has quit IRC (Remote host closed the connection) 2011-12-21T17:04:36 BenJackson: I'm not sure I'd put it that way. I would say that your combat has weaknesses that my combat exposes. I'm sure there are some bots that your combat will do better against than mine would. 2011-12-21T17:04:50 My combat has weaknesses too. 2011-12-21T17:05:05 _flag: iirc toLowerCase is called on the fill (true/false) strings to compare to 'true' 2011-12-21T17:05:52 jab_bott: yeah, double games like that are (were?) a bug due to a race condition in the scheduling 2011-12-21T17:06:06 *** lama3oid has quit IRC (Quit: Ðóñèôèöèðîâàííûé mIRC 6.35 îò yXo ** http://BestIRC.ru **) 2011-12-21T17:06:14 jab_bott: I don't know if they've got them all fixed yet, but I know they have been working on them. 2011-12-21T17:06:20 ChrisH_: ahh I see.. 2011-12-21T17:06:59 well it's clear from some games that the scoring and multi-player start positions sometimes create unfair games... but personally I think that's fine as on average over all the games it should even out 2011-12-21T17:08:14 <_flag> jab_bott: I think I've fixed it. It seems the canvas engine was either checking for or simply ignoring some of my input that was invalid, whereas the java version simply crashed on it, causing me to believe there was a problem with the java version; It works now 2011-12-21T17:08:34 My view is that "unfair starts" also reward bots that made an effort to play more strategically. 2011-12-21T17:08:35 *** amstan has joined #aichallenge 2011-12-21T17:08:35 *** ChanServ sets mode: +o amstan 2011-12-21T17:08:50 <_flag> jab_bott: Sorry about that 2011-12-21T17:09:14 _flag: no probs mate - was it the unclamped alphas? 2011-12-21T17:10:05 *** protocolocon has joined #aichallenge 2011-12-21T17:10:07 <_flag> jab_bott: Yeah, I didn't read carefully and I thought alpha was from 0 to 255 like everything else 2011-12-21T17:11:03 *** Varan has quit IRC (Quit: Leaving) 2011-12-21T17:11:33 _flag: yeah, there were a couple of inconsistencies from a graphics engine point of view, another one was the fact that most units were in map-coordinates, but line-width was in screen coordinates - probably the most sensible thing really, but still I didn't like that it wasn't consistent 2011-12-21T17:11:35 bugnuts: certainly a bot that can recover from an unfair start deserves extra credit 2011-12-21T17:12:48 *** Antimony has quit IRC (Ping timeout: 244 seconds) 2011-12-21T17:13:05 the leaderboard is starting to look more familiar 2011-12-21T17:13:37 oh.. I'm sure anyone who cares will be pleased to know I got 100% on ai-class 2011-12-21T17:13:39 *** Scryer has joined #aichallenge 2011-12-21T17:13:54 :< 2011-12-21T17:13:57 and now that we've got the cutoff up to 3000 the games are definitely more rapid for those above the cut 2011-12-21T17:14:05 jab_bott: nice job 2011-12-21T17:14:10 more rapid not exaclty 2011-12-21T17:14:23 jab_bott: i'm not pleased, lol 2011-12-21T17:14:27 jab_bott: i got way less than that 2011-12-21T17:14:30 *** jack_ has joined #aichallenge 2011-12-21T17:14:31 cut them to 1000 or 500 and they will be rapid 2011-12-21T17:14:44 I really wish I'd started learning ai a long time ago... there's so many things that have been solved that I've painstakingly done my own thing for in the past and had I known of these techniques would have made it so much easier 2011-12-21T17:15:02 I don't know about you, but a game once an hour is a huge improvement over what it was yesterday. 2011-12-21T17:15:18 *** observer has quit IRC (Quit: Page closed) 2011-12-21T17:15:37 amstan: well some of the questions were a bit didgy tbh.. but you have to take it in the context of the class and answer accordingly 2011-12-21T17:15:38 jab_bott: like what? 2011-12-21T17:15:50 ChrisH_ that it is 2011-12-21T17:15:50 jab_bott: what things? 2011-12-21T17:16:21 amstan: like A* - I never knew about A* before this ai-class.. even though I've done things similar (and spent ages tweaking them)... knowing about admissible heuristics would have saved me loads of time 2011-12-21T17:16:31 yeah, a* is cool 2011-12-21T17:17:45 *** murr4y has quit IRC (Read error: Connection reset by peer) 2011-12-21T17:19:01 ok, pull https://github.com/aichallenge/aichallenge/pull/391 is up 2011-12-21T17:19:09 amstan: also the stereo separation stuff... I wrote some code ages ago that generates those magic-eye (stereoscopic dot image things) images and I spent ages working out how to get the depths/offsets right 2011-12-21T17:20:39 amstan: my code in the end was a mess... it recursively shifted entire scanlines of repeating dots left/right to create different height layers 2011-12-21T17:21:33 *** jack_ has quit IRC (Quit: Page closed) 2011-12-21T17:22:47 I've spent so long reinventing stuff I didn't know existed already... like Sebastian said, it's better to learn about and use the existing knowledge so you can focus on pushing back the unknown a bit more 2011-12-21T17:23:13 _flag, hello, what did you use in battle calculation? 2011-12-21T17:24:20 avdg: tested? 2011-12-21T17:24:25 yeah 2011-12-21T17:24:55 *** ChrisH_ has quit IRC (Quit: Page closed) 2011-12-21T17:24:58 <_flag> teapotahedron_ni: Isolated groups of up to 8 or so ants (varied a bit depending on the situation), iterated through relevent possibilities and scored them 2011-12-21T17:25:34 damn i hate 10 player maps 2011-12-21T17:25:49 <_flag> I'll be explaining it more thoroughly in a couple days time when I get the chance, and I'll post a link in my profile for those who want to read about it :0 2011-12-21T17:25:52 <_flag> :)* 2011-12-21T17:27:18 thank you, it looks nice to me) 2011-12-21T17:27:36 <_flag> teapotaheron_ni: Thanks, you're doing pretty well too 2011-12-21T17:28:03 amstan: can the next challenge involve agents seeing noisy data and having to map the area themselves using particle systems? I realyl wnt to give that a go 2011-12-21T17:29:21 jab_bott: maybe, i really liked the asteroids idea 2011-12-21T17:30:15 Maybe a poll of the suggestions would give a good idea what idea is liked? 2011-12-21T17:30:30 amstan: assuming you're in the project root, just run "cd ants/visualizer && ant run-visualizer" 2011-12-21T17:31:15 *** Murashka has quit IRC (Quit: Page closed) 2011-12-21T17:31:19 amstan: I haven't read about that... but it gives me an idea... remember the megadrive (genesis) game SubTerrania? where you had a spaceship that could rotate left/right and thrust forwards/back, in a gravity environment... maybe something like that with asteroids too so you have to figure out the landscape and navigate it to rescue miners and things? 2011-12-21T17:33:05 http://aichallenge.org/visualizer.php?game=310286&user=12 2011-12-21T17:33:07 wtf 2011-12-21T17:33:13 my bot doesnt crash 2011-12-21T17:33:56 *** besh has joined #aichallenge 2011-12-21T17:34:05 * avdg wonders if the whole game could be "bugged" 2011-12-21T17:34:47 something smells here 2011-12-21T17:34:56 my bot should not have crashed 2011-12-21T17:35:03 and 2 timeouts? 2011-12-21T17:35:15 1 time outs regularly 2011-12-21T17:35:31 maybe it brought the us down 2011-12-21T17:35:35 -the 2011-12-21T17:35:38 jab bott: I used to have a stand-up video game of Gravitar, which is much like what you're describing. 2011-12-21T17:35:40 the other one I have to dig deeper (depending on ant count in other games) 2011-12-21T17:36:48 at least i hope people will crash/timeout evenly 2011-12-21T17:37:30 Sub-Terrania: http://www.youtube.com/watch?v=70XQsiTOK4I&feature=related Gravitar: http://www.youtube.com/watch?v=qIw6Mnw-NQY <= Yeah, something based on this kind of concept would be awesome! 2011-12-21T17:38:26 Timeouts/crashes are annoying, haven't had a crash in the final. But I've had several time-outs. 2011-12-21T17:38:27 maybe like a maze and you have to rescue/collect something and/or destroy somethign and escape/find the exit 2011-12-21T17:39:05 meh, it looks like the java visualizer also freezes my irc app (dunno, I have no hints) 2011-12-21T17:39:24 ... while the other players are rescuing and collecting stuff, or waiting to mug you for it at the exit? 2011-12-21T17:40:00 Scryer: perhaps.. maybe it's who can get there faster (no mugging) or who can be the most devious/strategic 2011-12-21T17:40:34 jab_bott: Blowing stuff up is engaging... that's why I suggested it. :) 2011-12-21T17:40:42 or even a game like marble madness would be cool, racing down some path with obstacles and trying to bash each othe roff the edge! :D 2011-12-21T17:42:14 jab_bott: see also "thrust" 2011-12-21T17:42:26 http://aichallenge.org/visualizer.php?game=310286&user=9885 I lose like 100 ants + at the exact same spot, i realize my fighting algo is so bad >< 2011-12-21T17:42:31 BenJackson: that archimedes game? heh we used to play that inschool all the time! 2011-12-21T17:42:40 I played on a C64 2011-12-21T17:42:59 jab_bott: also I pointed out xpilot 2011-12-21T17:43:09 BenJackson: oh.. was thinking of 'lander' 2011-12-21T17:43:25 mcstar : lol I just linked the same game oO 2011-12-21T17:43:27 *** Paradoxial has joined #aichallenge 2011-12-21T17:43:39 *** katebus has joined #aichallenge 2011-12-21T17:43:45 mcstar : thx for crashing btw (jk) 2011-12-21T17:43:51 http://www.youtube.com/watch?v=HOl2swTtheM 2011-12-21T17:44:01 skip in a bit for gameplay 2011-12-21T17:44:56 so nice 2011-12-21T17:46:09 *** epicmonkey has quit IRC (Ping timeout: 240 seconds) 2011-12-21T17:46:22 mcstar: that game ain't pretty 2011-12-21T17:46:33 no it's not :s 2011-12-21T17:46:39 I will only play if we also get the soundtrack... 2011-12-21T17:46:57 But Worldwar3 keeps timing out, so I guess it's not a server problem. I don't know about you mcstar 2011-12-21T17:47:17 dragonfrye12 times out at 200 bots 2011-12-21T17:47:35 im just surprised 2011-12-21T17:47:44 what language, mcstar ? 2011-12-21T17:47:44 my bot is fast and shouldnt have crashed 2011-12-21T17:47:53 this shitty something 2011-12-21T17:47:56 let me think 2011-12-21T17:47:59 BenJackson: What do you think of my performance when I fight between lakes? :) 2011-12-21T17:48:03 c++ + 11? 2011-12-21T17:48:07 I assume you've fed in the input from this game to a local copy 2011-12-21T17:48:15 you assumed wrong 2011-12-21T17:48:18 :D 2011-12-21T17:48:19 *** besh has quit IRC (Quit: Page closed) 2011-12-21T17:48:22 ill try 2011-12-21T17:48:32 Surya: hah, people forget it's water (and not walls). never heard "lakes 2011-12-21T17:48:52 *** amstan_ has joined #aichallenge 2011-12-21T17:48:52 *** ChanServ sets mode: +o amstan_ 2011-12-21T17:49:02 I see it from my ants point of view, it's water, it's cold. I don't know how to swim. 2011-12-21T17:49:03 that crashed bot with a 3-wide wall was a stout defender :) 2011-12-21T17:49:11 clearly 2011-12-21T17:49:13 It aren't pits? 2011-12-21T17:49:24 those 3 ants killed half of my ants 2011-12-21T17:49:26 my bot would have taken yellow 2011-12-21T17:49:37 mcstar: wellll 2011-12-21T17:49:44 you had an opportunity to take it after he timed out 2011-12-21T17:49:50 you got to his hill after he timed out 2011-12-21T17:49:53 I don't think so mcstar. You can't beat my fighting skills 2011-12-21T17:49:57 and he was still timed out and you didn't attack 2011-12-21T17:50:20 Surya: youre orange 2011-12-21T17:50:25 i meant worldwar3 2011-12-21T17:50:33 dark yellow 2011-12-21T17:50:38 many bots have the problem that the require movement of other ants to "jostle" them sometimes 2011-12-21T17:50:43 after that, id taken you :D 2011-12-21T17:50:51 ahaha :) 2011-12-21T17:51:19 youre right under me in ranking 2011-12-21T17:51:22 It's quite likely to be honest 2011-12-21T17:51:44 what ranking is that? 2011-12-21T17:51:51 around 180 ~ 2011-12-21T17:51:58 *** amstan has quit IRC (Ping timeout: 255 seconds) 2011-12-21T17:51:59 jostle? 2011-12-21T17:52:00 183 me 184 Surya 2011-12-21T17:52:12 I don't deserve 180 though. Aiming top 250 2011-12-21T17:52:16 bretep: other ants moving causing your ants to react differently 2011-12-21T17:52:25 some ants when confronted with stationary ants also become stationary 2011-12-21T17:52:47 *** Kettling has joined #aichallenge 2011-12-21T17:53:20 It's down to the point where I only win if my opponents time out... ha ha, at least I can write efficient code 2011-12-21T17:53:23 my ants detect standing ants 2011-12-21T17:54:40 *** TheLinker has joined #aichallenge 2011-12-21T17:55:00 CRAZY 2011-12-21T17:55:05 damn opera 2011-12-21T17:55:07 meh, my bot is on another losing streak 2011-12-21T17:55:21 in the titlebar the whole downloaded stream appeared 2011-12-21T17:55:23 In Game: Playing in a game right now. < oh crap 2011-12-21T17:55:32 i see that I have changed one param... from infinite to 3 times visibility radius to improve performance, but it changed the behaviour of ants so much, that from about 500 my bot now is in about 1000, 2011-12-21T17:55:56 katebus: mine has untested "improvements" too, damn 2011-12-21T17:56:06 ;-) anyway, this is fun, this was fun, 2011-12-21T17:56:12 but crashing wasant one of them... 2011-12-21T17:56:28 I learned how to use git, I learned pythong (after seeing so many nice examples on forum and in tutorial I decided to learn) 2011-12-21T17:56:43 so many benefits 2011-12-21T17:57:07 it is just a pity, that I have now a bot, that is not much better than GreedyBot 2011-12-21T17:57:24 i left the pipe out and cat'd my bot 2011-12-21T17:57:35 poor bot got smashed onto the screen 2011-12-21T17:57:41 mcstar: what do you say about crashing 2011-12-21T17:57:53 you keep saying pythong 2011-12-21T17:58:06 ha ha ha pythong ;-) 2011-12-21T17:58:07 ythong>python 2011-12-21T17:58:13 trying out the input from the vis. 2011-12-21T17:58:14 python* 2011-12-21T17:58:27 my new keyboard ;-) cant get used to 2011-12-21T17:58:28 pythong is pretty good 2011-12-21T17:58:35 i cannot believe my luck: http://aichallenge.org/visualizer.php?game=310531&user=5916 with the last ant! 2011-12-21T17:58:46 (bad luck) 2011-12-21T17:59:12 protocolocon: step 1, click your link, step 2, quickly scan the list for my name in case I will be the victim, step 3, relax 2011-12-21T17:59:14 pythong fits women well 2011-12-21T17:59:34 the old one (keyboard) was fighting with the orange juice, and the orange juice won the fight 2011-12-21T17:59:51 time cat crash.txt | pattern/MyBot > out 2011-12-21T17:59:52 mcstar: ;-) 2011-12-21T17:59:53 Segmentation fault 2011-12-21T17:59:55 real 0m21.516s 2011-12-21T17:59:57 user 0m21.372s 2011-12-21T17:59:59 sys 0m0.133s 2011-12-21T18:00:00 http://aichallenge.org/visualizer.php?game=310577&user=2255 I almost didn't lose 2011-12-21T18:00:15 thestinger: is that an input to trueskill? 2011-12-21T18:00:15 oh so you have segfault 2011-12-21T18:00:38 *** bearoff has left #aichallenge 2011-12-21T18:00:42 protocolocon: to be fair you chased him there 2011-12-21T18:00:44 :) 2011-12-21T18:00:50 not very nice, doesn't look like a planned feature 2011-12-21T18:00:55 pguillory stole my kill of yellow's hill :( 2011-12-21T18:00:59 my bot would probably do the same thing 2011-12-21T18:01:04 but thats 54ms per turn 2011-12-21T18:01:07 so im at least fast 2011-12-21T18:01:47 thestinger: the breakout and recaptuer at 770 is awesome! 2011-12-21T18:01:55 *** thagomizr has quit IRC (Quit: thagomizr) 2011-12-21T18:02:48 Crashing at high speeds is very bad for your health 2011-12-21T18:02:54 he's like a spider "I'll eat you later" 2011-12-21T18:03:30 probably its gcc's fault 2011-12-21T18:03:30 orange really crushed my offense 2011-12-21T18:03:49 it must have misunderstood me, and compiled something in a bad way 2011-12-21T18:03:53 he retreats with his lower half so my ants go into a ball 2011-12-21T18:04:20 mcstar: on x86 and x86_64 it's usually not gcc's fault... 2011-12-21T18:04:37 BenJackson: im just trying clear my name 2011-12-21T18:04:43 It's a glitch in the cpu 2011-12-21T18:04:46 thestinger: I considered looking for ways to force combat to put enemies against a wall but i never did 2011-12-21T18:04:55 mcstar: I wouldn't discover that without you pointing to my typos in "python" -> http://pythong.org/ 2011-12-21T18:05:08 *** olexs1 has joined #aichallenge 2011-12-21T18:05:10 it seems combat abily is very much a function of curvature 2011-12-21T18:05:14 ability 2011-12-21T18:05:35 I appreciate pythong less after seeing that link 2011-12-21T18:05:41 katebus: lol :S 2011-12-21T18:05:42 my bot is good at combat when the enemy cooperates :) 2011-12-21T18:06:06 seems its a common typo 2011-12-21T18:06:13 pguillory and xathis do combat in a way that my bot can deal with 2011-12-21T18:06:13 even Mics gave his name 2011-12-21T18:06:18 (we call him Mics) 2011-12-21T18:06:19 My bot is also good in combat if the enemy sends the ants 1 for 1 2011-12-21T18:06:25 s/xathis/protocolocon/ 2011-12-21T18:06:40 _flag, good game, thanks for playing) ~60 ants ahead... 2011-12-21T18:06:55 omg, im running my bot under valgrind since i did the paste 2011-12-21T18:06:57 but cheeser does a fancy encircling maneuver 2011-12-21T18:06:59 it isnt finished 2011-12-21T18:07:05 thestinger: isn't every bot good in combat when enemy cooperates? 2011-12-21T18:07:10 Stack overflow in thread 1: can't grow stack to 0x7fe801ff8 2011-12-21T18:07:17 katebus: well, pguillory plays by the rules my bot expects 2011-12-21T18:07:23 mcstar: infinite recursion? 2011-12-21T18:07:37 aaaaa my gonnaStay function recurtses 2011-12-21T18:07:41 katebus: so good, cautious combat == cooperating :) 2011-12-21T18:07:41 damn it 2011-12-21T18:07:42 <_flag> teapotahedron_ni: Good game indeed, looks like it may be a close battle for second :) 2011-12-21T18:07:44 *** olexs has quit IRC (Ping timeout: 252 seconds) 2011-12-21T18:07:46 i thought i got rid of it 2011-12-21T18:07:49 katebus: ok is it "kah-TEE-bus" or "KATE-bus" 2011-12-21T18:07:54 katebus: good, aggressive combat destroys me 2011-12-21T18:08:04 BenJackson: its a checker, to avoid stepping on each other 2011-12-21T18:08:13 it recursively deletes orders back 2011-12-21T18:08:20 mcstar: loop of stepping-on then? 2011-12-21T18:08:28 I've noticed that even the best bots don't seem to pile in even when they could easily win 2011-12-21T18:08:38 <_flag> teapotaherdon_ni: Are you a team of three people? 2011-12-21T18:08:46 i dunno 2011-12-21T18:08:51 I guess I optimized too much against the opponents on the tcp servers 2011-12-21T18:08:57 _flag: i hope it will be 2011-12-21T18:09:16 *** meduza has quit IRC (Quit: Page closed) 2011-12-21T18:09:22 jab_bott: I think with more time we could have a new top 10 of bots that would destroy xathis and the rest of the current top 10 2011-12-21T18:09:23 *** olexs1 has quit IRC (Ping timeout: 252 seconds) 2011-12-21T18:09:25 i cant imagine a loop condition arising 2011-12-21T18:09:30 BenJackson: well it is a stupid nick - the katebus - I don't even remember how it was created, I have to find something beter, but the first one is better 2011-12-21T18:09:33 (but it does) 2011-12-21T18:09:47 mcstar: 4 bots in a square moving in a circle plus a new bot trying ot enter the circle 2011-12-21T18:09:50 _flag: yes, we are =) 2011-12-21T18:10:01 *** Storm__ has quit IRC (Quit: Page closed) 2011-12-21T18:10:04 BenJackson: no 2011-12-21T18:10:11 it deletes pointers from a 2d map 2011-12-21T18:10:13 *** McLeopold has joined #aichallenge 2011-12-21T18:10:20 *** gaussin has left #aichallenge 2011-12-21T18:10:25 katebus: there are a lot of people from all over the world here so I thought maybe it was something I just didn't recognize 2011-12-21T18:10:27 it deletes first, and recurses afterwards 2011-12-21T18:10:32 _flag: yes 2011-12-21T18:10:51 BenJackson: do you really think that xathis cannot use alle the time given? 2011-12-21T18:11:14 well, we will never know, until we try 2011-12-21T18:11:20 nobody (afaik) is truly looking ahead multiple turns 2011-12-21T18:11:30 there are lots of visible combat mistakes on that scale 2011-12-21T18:11:36 https://github.com/liquid-phynix/aicbot/blob/master/base.cpp#L395 2011-12-21T18:11:45 seems it was just an idea in my head 2011-12-21T18:11:46 *** jab_bott has quit IRC (Quit: Page closed) 2011-12-21T18:11:49 that i didnt follow 2011-12-21T18:11:49 most of the strength estimations are pretty iffy 2011-12-21T18:11:56 only ChrisH and one other guy tried hill control 2011-12-21T18:12:12 lol, my strength estimation for battles was a last minute addition 2011-12-21T18:12:16 oh, sure, you are right looking ahead, checking the decision tree... too much computaion for a short time limit 2011-12-21T18:12:27 and the line formation hints 2011-12-21T18:12:34 all in the last 2 hours :) 2011-12-21T18:12:40 *** olexs has joined #aichallenge 2011-12-21T18:12:43 katebus: maybe, but maybe there are more efficient data structuers 2011-12-21T18:13:03 also nobody is really saving info from turn to turn 2011-12-21T18:13:06 which has pros and cons 2011-12-21T18:13:13 so maybe I can't look ahead 5 turns in one turn 2011-12-21T18:13:18 it complicated matters too much for me 2011-12-21T18:13:22 i couldnt hsndle the complexity 2011-12-21T18:13:22 but maybe I can look ahead 2 turns per turn and build up 20 turns ahead over 10 turns 2011-12-21T18:13:25 for a strategic goal 2011-12-21T18:14:53 katebus: plus there are lots of areas where you could invest a lot of effort for a small gain 2011-12-21T18:14:59 eg symmetry isn't a huge win, but it's something 2011-12-21T18:15:01 *** tomb___ has joined #aichallenge 2011-12-21T18:15:10 http://sprunge.us/HRVh that was the only half-intelligent thing I included in my bot... 2011-12-21T18:15:15 sure, it is all about good heuristic, or just a trick 2011-12-21T18:15:15 or a near-perfect travelling salesman solution to the initial food 2011-12-21T18:16:17 does Accoun still want people to play on his server? 2011-12-21T18:16:31 gn8 everyone! 2011-12-21T18:16:41 thestinger: https://github.com/bjj/ants2011/blob/master/combat.cc#L466 2011-12-21T18:16:56 *** olexs1 has joined #aichallenge 2011-12-21T18:17:04 BenJackson: I have to do something similar 2011-12-21T18:17:05 *** olexs has quit IRC (Ping timeout: 252 seconds) 2011-12-21T18:17:14 *** Surya has quit IRC (Quit: Page closed) 2011-12-21T18:17:32 otherwise my ants will never move onto a hill if an ant is on it at the start of the turn (even if I predict they will move it off) 2011-12-21T18:17:37 *** bretep has quit IRC (Ping timeout: 258 seconds) 2011-12-21T18:18:23 thestinger: that code is duplicated in my territory code but it's sort of meaningless sicne if there IS an enemy ant it willb e in combat 2011-12-21T18:19:53 people are saying good night, that reminds me I am a human not a bot, so I need to sleep also 2011-12-21T18:20:09 so good night everyone 2011-12-21T18:20:17 cya 2011-12-21T18:21:25 *** katebus has quit IRC (Quit: Page closed) 2011-12-21T18:21:38 by kate 2011-12-21T18:22:31 *** Akranis has quit IRC (Read error: Connection reset by peer) 2011-12-21T18:23:18 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-12-21T18:25:29 BenJackson, and has anyone made any good approximation to salesman problem solution? 2011-12-21T18:25:47 for ants? 2011-12-21T18:25:59 yes) 2011-12-21T18:26:03 for food search 2011-12-21T18:26:06 not that I know of 2011-12-21T18:26:21 tricky because you want to consider ants that appear at your hills 2011-12-21T18:26:35 It seems to me overkill to do it that way: in classical TSP the salesmen can't be shot. 2011-12-21T18:27:01 hmmm, we actually do ;) but is that helping... i don't know 2011-12-21T18:27:14 Scryer: for the very very beginning 2011-12-21T18:27:16 not all the time 2011-12-21T18:27:21 we created virtual ants in hills in case of 1 hill 2011-12-21T18:27:27 *** keith_ has joined #aichallenge 2011-12-21T18:27:38 *** olexs has joined #aichallenge 2011-12-21T18:28:02 invasion of the hedraas 2011-12-21T18:28:52 how did you guys cooperate? 2011-12-21T18:28:58 in person? 2011-12-21T18:29:09 mcstar, mostly in person 2011-12-21T18:29:19 must have been fun 2011-12-21T18:29:27 but you have to share the credit 2011-12-21T18:29:44 *** olexs1 has quit IRC (Ping timeout: 252 seconds) 2011-12-21T18:29:59 it would be a fun problem to solve 2011-12-21T18:30:08 that's ok for us) we shared the fun 2011-12-21T18:30:11 but like I was saying, given enough time we could make a very sophisticated bot 2011-12-21T18:30:14 it was alot of fun, nights and days 2011-12-21T18:30:18 but was that worth focusing on during the contest? hard to say 2011-12-21T18:31:18 Tetrahedron i wonder what your thoughts are on my. Path finding example. The other day 2011-12-21T18:31:30 But I don't have the pStebin link handy 2011-12-21T18:32:38 I better finish better than 36th 2011-12-21T18:32:45 Antimatroid: bfs? =) 2011-12-21T18:33:05 Nah, it. Wasn't that simple 2011-12-21T18:33:38 BenJackson how easily could you finds that paste bin example I had for you? 2011-12-21T18:34:15 Sorry about my typing, I'm on someone else's iPad which i don't know how to use very well 2011-12-21T18:35:19 Antimatroid: I don't really know how to drive irssi very well 2011-12-21T18:35:19 BenJackson: Use weechat. 2011-12-21T18:35:40 after the challenge I will resume not using IRC 2011-12-21T18:35:46 irssi 2011-12-21T18:35:46 mcstar: Use weechat. 2011-12-21T18:35:52 xchat 2011-12-21T18:35:59 bitchx 2011-12-21T18:36:02 irssi 2011-12-21T18:36:02 mcstar: Use weechat. 2011-12-21T18:36:06 hm 2011-12-21T18:36:09 irssi 2011-12-21T18:36:09 amstan_: Use weechat. 2011-12-21T18:36:41 ZenWraithBot: help 2011-12-21T18:36:44 !timeleft 2011-12-21T18:36:46 !help 2011-12-21T18:36:59 k, i'll handle it later 2011-12-21T18:37:05 :) 2011-12-21T18:37:56 *** Zedenstein has quit IRC (Ping timeout: 258 seconds) 2011-12-21T18:42:53 someone built an irssi bot? ;) 2011-12-21T18:42:54 Zannick: Use weechat. 2011-12-21T18:43:13 the opposite 2011-12-21T18:43:16 weechat 2011-12-21T18:43:21 *** ChanServ sets mode: +o Zannick 2011-12-21T18:43:23 *** Zannick sets mode: +q ZenWraithBot!*@* 2011-12-21T18:43:26 Use irssi. 2011-12-21T18:43:31 :D 2011-12-21T18:43:37 ZenWraithBot: use irrsi. 2011-12-21T18:43:43 *** Zannick sets mode: -o Zannick 2011-12-21T18:43:51 *** Scryer has quit IRC (Remote host closed the connection) 2011-12-21T18:45:37 aha 2011-12-21T18:45:38 *** gaussin1 has joined #aichallenge 2011-12-21T18:45:43 i've found the original purpose 2011-12-21T18:45:45 *** ChanServ sets mode: +o Zannick 2011-12-21T18:45:50 *** Zannick sets mode: -q ZenWraithBot!*@* 2011-12-21T18:45:52 !timeleft 2011-12-21T18:46:07 no? 2011-12-21T18:46:09 irssi? 2011-12-21T18:46:10 Zannick: Use weechat. 2011-12-21T18:46:27 *** Zannick sets mode: +q ZenWraithBot!*@* 2011-12-21T18:46:30 never mind then. 2011-12-21T18:46:33 *** Zannick sets mode: -o Zannick 2011-12-21T18:48:50 use irclib to write your own irc client 2011-12-21T18:50:12 *** Antimatroid has quit IRC (Ping timeout: 258 seconds) 2011-12-21T18:51:27 http://aichallenge.org/visualizer.php?game=310926&user=4513 2011-12-21T18:51:46 it's clearly not a bug 2011-12-21T18:52:03 A good feature would be if the players in a game could add their comments to the bottom of that game (like a mini chat room) 2011-12-21T18:53:04 I think that would encourage "The Spirit of the Game" kind of stuff 2011-12-21T18:53:10 yes, we also thought about that) and i think i remember a forum post about it. it would be great 2011-12-21T18:53:50 Doesn't PHP support chat-room as a builtin command? 2011-12-21T18:55:57 what? haha. 2011-12-21T18:56:01 and it would also be cool to rate some of the games 2011-12-21T18:56:37 +) 2011-12-21T18:56:48 like on facebook lol 2011-12-21T18:57:11 *** Antimatroi has joined #aichallenge 2011-12-21T19:00:20 Green tea is rising 2011-12-21T19:00:56 *** protocolocon has quit IRC (Ping timeout: 258 seconds) 2011-12-21T19:01:40 And pguillory rose 6 places 2011-12-21T19:02:47 jcdny: you around? 2011-12-21T19:02:50 pairofdice: by stealing my kill! :P 2011-12-21T19:03:13 Thestinger are you surprised? D: 2011-12-21T19:03:31 That was meant do be d: 2011-12-21T19:03:35 http://aichallenge.org/visualizer.php?game=310577&user=589&turn=398&row=48&col=88 2011-12-21T19:04:15 That doesn't work well on this 2011-12-21T19:04:23 *** QGazQ has quit IRC (Quit: sleep) 2011-12-21T19:05:55 Im in a game, that is both exciting and terrifying 2011-12-21T19:07:43 I'm in a game too :0 2011-12-21T19:08:27 McLeopold: yes 2011-12-21T19:08:52 bugnuts: I've got a symmetry question. What do you mean by cycle length? 2011-12-21T19:09:10 number of points which are in the equivalence class. 2011-12-21T19:09:42 so if you tile 2 squares which have rotational symmetry it would be 8 2011-12-21T19:09:44 so, what's that? 2011-12-21T19:10:11 Thestinger I reckon we're playing each other 2011-12-21T19:10:16 bugnuts: you lost me 2011-12-21T19:10:23 <_flag> McLeopold: What is the plan for the size of the final cutoff? 1000 ants? 100? 2011-12-21T19:10:30 500 2011-12-21T19:10:37 janzert said 500, maybe 250 if needed 2011-12-21T19:10:50 minus FlagCapper 2011-12-21T19:11:05 <_flag> McLeopold: Me winning will be a forgone conclusion then? 2011-12-21T19:11:10 <_flag> :0 2011-12-21T19:11:13 <_flag> :) 2011-12-21T19:11:15 there are definitely still starter bots circa 3000 2011-12-21T19:11:15 McLeopold symmetrically equivalent locations are in the same equivalence class and the equivalence classes partition the map 2011-12-21T19:11:58 well for a single mirror every point has 1 other point that is equivalent. a second mirroring doubles that to 4. Then if you take that rectangle with 4 points that are the same and tile the torus 2011-12-21T19:12:11 with 6 of them it would then be 24 2011-12-21T19:12:18 i don't think any of the map symmetries are that complicated though 2011-12-21T19:12:29 but i guess they could have been, and you would have detected it, so bravo 2011-12-21T19:12:43 I did the symmetries 2011-12-21T19:12:48 there are ones which are 1 mirroring that is then tiled. 2011-12-21T19:13:06 Bug nuts, did translating the rotationally symmetric maps make detection harder for you? 2011-12-21T19:13:12 Antimony said not really 2011-12-21T19:13:16 not at all. 2011-12-21T19:13:30 i hate to say this but 2011-12-21T19:13:35 It did at least rule out trivial symmetry detection 2011-12-21T19:13:35 FUCKING TIMEOUTS 2011-12-21T19:13:43 they dont happen locally 2011-12-21T19:13:59 im getting frustrated 2011-12-21T19:14:00 flowenol: i just catched a stack overflow in mine 2011-12-21T19:14:00 the ones that are hard are the ones with "almost symmetry" 2011-12-21T19:14:02 flowenol: do you produce any/a lot of output on stderr? 2011-12-21T19:14:02 Flowenol: you need to test these things beforehand 2011-12-21T19:14:10 a1k0n none 2011-12-21T19:14:18 i thought it coudldnt happen too :S 2011-12-21T19:14:36 well it should be repeatable 2011-12-21T19:14:37 Antimatroi: didn't you find the example? 2011-12-21T19:14:42 and they are not 2011-12-21T19:14:46 bugnuts: does symmetry detection help your ant? 2011-12-21T19:15:11 *** replore_ has joined #aichallenge 2011-12-21T19:15:15 argh its depressing 2011-12-21T19:15:41 Teapotahedron_ni: I can't on here, can someone else search the logs for a paste bin link from. Me a few days ago? 2011-12-21T19:16:27 I think it helps a fair bit with food gathering on sparser maps. 2011-12-21T19:16:32 Antimatroi, i think i can... don't you remember the date? 2011-12-21T19:16:34 bugnuts: how does tiling 2 squares produce 8 symmetric points? wouldn't it be 2 sym points? 2011-12-21T19:16:39 *** GreenTea has quit IRC (Ping timeout: 258 seconds) 2011-12-21T19:16:50 flowenol: with a language like C# or Java I wouldn't say timeouts (say due to GC) would be repeatable without the same runtime and runtime settings 2011-12-21T19:16:57 tiling a square that has rotational symmetry gives you 8 2011-12-21T19:17:17 BenJackson tbh thats the only thing left to blame 2011-12-21T19:17:27 and i always forget that one 2011-12-21T19:17:38 For those asking the next cutoff change will probably not be for another 10-12 hours from now at least 2011-12-21T19:17:45 :( 2011-12-21T19:17:47 bugnuts: I don't get it? 2011-12-21T19:17:47 basically you can take any rectangle with some symmetry and tile a bigger torus with it... 2011-12-21T19:17:53 *** Antimatroi has quit IRC (Quit: Page closed) 2011-12-21T19:17:55 janzert at what value? 2011-12-21T19:18:18 next stop is either 1500 or 2000 probably the latter 2011-12-21T19:18:19 okay, 1 square with rotational sym has 2 sym points? so create 1 copy tile, and you have 4 2011-12-21T19:18:25 *** Antimatroid has joined #aichallenge 2011-12-21T19:18:43 janzert ok sounds good to me :D 2011-12-21T19:18:50 im at 452 2011-12-21T19:19:04 equivalency points 2011-12-21T19:19:09 *** AlliedEnvy has quit IRC (Read error: Connection reset by peer) 2011-12-21T19:19:35 *** Zedenstein has joined #aichallenge 2011-12-21T19:19:58 *** FranPeruzzi has joined #aichallenge 2011-12-21T19:20:09 *** AlliedEnvy has joined #aichallenge 2011-12-21T19:20:11 ok, my sleep schedule is completely off this week since dealing with the finals and atm am at the up for 30+ hours point so I'm heading for the down for 10+ hours time. I'll see everyone tomorrow :) 2011-12-21T19:20:20 think about it like this - if you had a 60x60 map with rotational symmetry you would have 4 pouints that were the same. If you just pasted it to itself you would have a 60x120 map with 8 points 2011-12-21T19:20:22 thanks janzert 2011-12-21T19:20:24 unless there's a last minute need right now for me :) 2011-12-21T19:21:01 bugnuts: okay, you are thinking rotational symmetry is a central point mirrored vert and horz? 2011-12-21T19:21:20 *** antimac has joined #aichallenge 2011-12-21T19:21:23 yeah 2011-12-21T19:21:49 Antimatroid: this? http://pastebin.com/imX2uBM2 2011-12-21T19:22:06 bugnuts: so, for cycle length of translational symmetry, are you calculating 1 number or 2? 2011-12-21T19:22:35 for translation it is the number of points until you return to the same point. 2011-12-21T19:23:02 tetrahedron: yeah 2011-12-21T19:23:07 was just about to paste it to you :P 2011-12-21T19:23:10 going which direction? with map wrapping? 2011-12-21T19:23:42 it does not matter which direction but with wrapping yeah. 2011-12-21T19:23:43 tetrahedron: problem: with a single bfs from *'s move all ants optimally towards them 2011-12-21T19:24:00 ignore whether that makes sense in the game 2011-12-21T19:24:06 just treat that as the problem 2011-12-21T19:24:53 bugnuts: so, 2 numbers? 2011-12-21T19:25:11 bottom right ant should walk south and around the line, ant ant in line should step out and around to let the "trapped ant" in, and the ant with a gap below it should walk south (so you can't block path finding with ants) 2011-12-21T19:25:22 an* ant in line 2011-12-21T19:25:44 no one number 2011-12-21T19:25:52 also, _flag lost to GreenTea with 1v1 2011-12-21T19:26:51 you have the number of steps to get back to the same row and steps to get back to the same col and the total steps is the least common multiple of those 2 numbers 2011-12-21T19:26:53 bugnuts: okay, I may be stupid, but if you walked north, you'll hit the same point in the same time as the number of rows, for random walk maps 2011-12-21T19:27:09 *** AlliedEnvy has quit IRC (Read error: Connection reset by peer) 2011-12-21T19:27:13 tetrahedron: got any ideas of how to accomplish all of that with one bfs? :P 2011-12-21T19:27:38 the hardest is having the bottom right ant step around while the ant with a gap steps down 2011-12-21T19:27:57 ah no, you first identify 2 points that have the same neighborhood. the difference between them defines an offset in rows/cols 2011-12-21T19:28:19 *** AlliedEnvy has joined #aichallenge 2011-12-21T19:28:47 *** - do these stars mean food? 2011-12-21T19:28:52 on the map 2011-12-21T19:29:05 ignore their relevance to the game 2011-12-21T19:29:15 they are targets which you want to move all ants towards optimally 2011-12-21T19:29:24 make them enemy hills if that helps 2011-12-21T19:29:32 you then add that offset repeatedly to p1 wrapping as needed and the number of additions is the cycle length. 2011-12-21T19:29:47 sorry number of additions to return to p1. 2011-12-21T19:29:54 ah 2011-12-21T19:30:07 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-12-21T19:30:46 but if the offset was {r,c} and the map size was {R,C} its also LCM(LCM(r,R),LCM(c,C)) 2011-12-21T19:32:40 ok, it's not too hard to launch bfs from each of our targets and consider some ants "busy" when we find them. still no idea how to do it correct with a singe bfs 2011-12-21T19:33:50 antimac, and what does optimally mean here? 2011-12-21T19:33:52 note do self: dont use java next time 2011-12-21T19:34:09 C++ all the way 2011-12-21T19:34:10 yeah, it's the single bfs part that stumped me, i was trying to work out how to optimally route ants when i am just moving all ants at the end to their closest target 2011-12-21T19:34:15 flowenol, why not? 2011-12-21T19:34:30 teapotahedron_ni first, I dont like java 2011-12-21T19:34:32 second GC 2011-12-21T19:34:35 well 2011-12-21T19:34:41 :) 2011-12-21T19:34:46 its the only thing that can be making me timeout atm 2011-12-21T19:34:59 we had an idea for idle ants which were not used for anything 2011-12-21T19:35:24 at least with C++ ill manage the memory myself 2011-12-21T19:35:29 we would send them to the next cell which is nearer to the aim 2011-12-21T19:35:35 with 1 bfs 2011-12-21T19:35:42 maybe it could help here? 2011-12-21T19:35:54 Re-use your structures and the GC problem goes away 2011-12-21T19:36:02 teapotahedron_ni: http://pastebin.com/Gp2hq9Dh you want to get that 2011-12-21T19:36:06 flowenol: use c++11 and magic 2011-12-21T19:36:23 im way too behind C++ specs 2011-12-21T19:36:32 care to enlighten me? 2011-12-21T19:36:40 finally 2011-12-21T19:36:51 i've got what's the problem is) 2011-12-21T19:36:58 it's worth noting that if you search over friends but don't move ants statically, then sometimes the bottom right ant might move east instead of south, because they both appear to have the same path distance to the targets 2011-12-21T19:37:10 *** Jak_o_Shadows1 has joined #aichallenge 2011-12-21T19:38:41 *** gaussin1 has quit IRC (Quit: Leaving.) 2011-12-21T19:39:02 * antimac steps in front of thestinger 2011-12-21T19:39:19 *** Jak_o_Shadows has quit IRC (Ping timeout: 252 seconds) 2011-12-21T19:39:58 hmm, semi bad example with the gap, because the bottom right 'a' could optimally wait for the gap, but you get what i mean 2011-12-21T19:40:12 *** mcstar has quit IRC (Quit: WeeChat 0.3.6) 2011-12-21T19:40:13 the gap was added to highlight that the line needs to continue moving south 2011-12-21T19:40:26 still don't know, what to do with it. maybe, something like a1k0n presented should work. if we calculate value function for each cell and enable tradeoff between cells, we can somehow get the solution - we make one ant stay worse, but a bunch of other ants stay better. am I correct? 2011-12-21T19:40:37 *** TheLinker has quit IRC (Remote host closed the connection) 2011-12-21T19:40:45 antimac: can't you just not expand an ant node? 2011-12-21T19:40:55 McLeopold: not optimal 2011-12-21T19:41:24 i compromised, what i did was.. 2011-12-21T19:41:31 or you could try to solve this problem with hungarian algorithm maybe 2011-12-21T19:42:15 search over ants and move towards a closest target if possible, otherwise don't make a move, so bottom right 'a' would sit still and trapped 'a' would walk down until it was fully trapped 2011-12-21T19:42:25 then to help get trapped ants out i would randomise path finding 2011-12-21T19:43:06 although if bottom right 'a' was then on the shortest path to another target it would go off in that direction 2011-12-21T19:43:10 antimac, do you mean randomizing just the order of directions? 2011-12-21T19:43:22 which leads to ants jumping around a bit, but also alows ants to nicely "flow" paste targets that block up 2011-12-21T19:43:38 teapotahedron_ni: yeah, so it sort of does a random walk along the shortest paths 2011-12-21T19:43:47 although you can have chokepoints that all shortest paths go through 2011-12-21T19:43:53 antimac: I think it would be optimal, but I need to test it now :) 2011-12-21T19:43:54 which is there the ants block up 2011-12-21T19:44:01 where* 2011-12-21T19:44:21 *** grwip has quit IRC (Quit: Leaving) 2011-12-21T19:44:34 one can also randomize pathfindig to find not optimal but suboptimal solutions... may be interesting 2011-12-21T19:44:40 randomising shortest paths isn't as simple as i thought either 2011-12-21T19:44:47 i do it the trivial way, but it's far from optimao 2011-12-21T19:46:19 you know to walk randomly away from optimal direction. it's mostly useless, but in case of many ants it could help them to spread 2011-12-21T19:47:21 but if the ants are blocked to their target and it stays blocked, it allows the ants to flow past the blockage 2011-12-21T19:47:47 so i only move ants statically if they're on the battle border, otherwise an ant doesn't specifically sit still, so they can move through blockages 2011-12-21T19:47:59 my bot isn't winning, so i thought that was cool and left it 2011-12-21T19:48:16 not through, past/around 2011-12-21T19:49:04 do they acctually walk around blockages? 2011-12-21T19:49:35 *** danapplegate has quit IRC (Quit: Ex-Chat) 2011-12-21T19:49:39 not in the blockage as given in the example above 2011-12-21T19:49:45 antimac: lol, another long streak of losses... 2011-12-21T19:50:06 my stupid bot >.< 2011-12-21T19:50:58 interesting problem in general 2011-12-21T19:51:33 http://pastebin.com/78F3qu8j 2011-12-21T19:51:38 take that example 2011-12-21T19:52:07 at the end you probably want to move the top row ants around towards the free battle border 2011-12-21T19:52:47 i think the previous example could be solved by randomly making "mistakes" in pathfinding. of course - it's not optimal, but.. 2011-12-21T19:53:44 and in the new example neither do i know, what to do. it's interesting, how we can surround enemy taking not the shortest paths 2011-12-21T19:54:17 *** JorgeB has quit IRC (Read error: Operation timed out) 2011-12-21T19:54:56 *** nplus has quit IRC (Quit: ChatZilla 0.9.87-rdmsoft [XULRunner 1.9.0.17/2009122204]) 2011-12-21T19:55:31 *** nplus has joined #aichallenge 2011-12-21T19:56:14 maybe, one can somehow identify the blocades and make that squares impassable 2011-12-21T19:56:23 *** valydo has joined #aichallenge 2011-12-21T19:57:48 that's what i do 2011-12-21T19:58:07 often if soonestBattle < 4 and a square is taken i'll treat it was blocked 2011-12-21T19:58:14 as* 2011-12-21T19:58:59 cool, we didn't actually do anything about it, so we can see stupid blocades 2011-12-21T19:59:08 yeah, we had a thought about labeling some of the cells as impassible 2011-12-21T19:59:25 by making bfs from water cells 2011-12-21T19:59:30 often then i don't leave enough backup around for when they break through the stand off though 2011-12-21T20:01:19 *** arscan has quit IRC (Quit: Page closed) 2011-12-21T20:02:36 yep. i think very intresting would be something like "psycology analizer" for enemy behaviour. Something like: "we see exchanges -> enemy is getting aggressive in that area -> we need backup" 2011-12-21T20:03:12 antimac: your right, I'd have to make a change to my bfs for the bottom ant 2011-12-21T20:03:29 McLeopold: you will walk around in circles if you try and fix things :) 2011-12-21T20:03:32 I could unexpand a node if an ant moved there... 2011-12-21T20:03:43 you'll fix one then, think everything is good, then notice you broke another thing 2011-12-21T20:04:04 but that doesn't solve the problem in the top section, unless the left ant is chosen first 2011-12-21T20:04:06 i am very interested if anyone can actually solve that problem though 2011-12-21T20:04:57 *** raja_ has quit IRC (Ping timeout: 258 seconds) 2011-12-21T20:05:23 I use a python generator, so at the point I expand the spot where 2 ants want to go, I could schedule an ant, then stop expanding 2011-12-21T20:05:38 I can then pass back to the generator to stop expanding that current node 2011-12-21T20:05:53 but then, the upper ant may be chosen last, and would go left 2011-12-21T20:06:13 McLeopold: one option is to just allow the ant to be trapped (move him south for the moment), then randomised path finding will eventually get him out 2011-12-21T20:06:17 *** codetiger has joined #aichallenge 2011-12-21T20:06:20 it's not optimal but not bad 2011-12-21T20:06:59 to solve, I would have to mark to mark both ants as moving to the same spot, continue expanding, and pick the first location that either ant can go to to resolve 2011-12-21T20:07:56 for the upper part, I'm thinking you can delay the solution until the ant hits the wall, in which case it would only have 1 direction, and be chosen as the ant to move into the contested spot 2011-12-21T20:08:18 still all doable with 1 bfs 2011-12-21T20:09:29 *** LouisMartin has quit IRC (Quit: Page closed) 2011-12-21T20:09:47 so, basically, any spot where multiple ants want to go, solve the solution where the total distance for all of them is minimal 2011-12-21T20:10:11 and if 1 of the ants only has 1 direction, that takes priority 2011-12-21T20:11:49 *** Jak_o_Shadows1 has quit IRC (Remote host closed the connection) 2011-12-21T20:13:51 *** Redgis has quit IRC (Ping timeout: 252 seconds) 2011-12-21T20:16:11 that could work 2011-12-21T20:16:15 i have to go out for a while 2011-12-21T20:16:28 any plans to lower cutoff ahead of 3000 2011-12-21T20:16:43 in about 10+ horus 2011-12-21T20:16:51 hours even 2011-12-21T20:17:05 seems like a long time. is there still churn in the 2-3000 range? 2011-12-21T20:17:08 *** Migi32 has quit IRC (Remote host closed the connection) 2011-12-21T20:17:31 we are converging quite nicely now http://pastebin.com/qyAhD1dd 2011-12-21T20:18:12 I think we can get all sigmas below 1.5 in the next 2 days 2011-12-21T20:18:17 how is top by mu tracking top by score? 2011-12-21T20:18:30 skill? 2011-12-21T20:18:37 er, yes 2011-12-21T20:18:44 let me check 2011-12-21T20:20:04 my sigma is on the high end 2011-12-21T20:20:09 http://pastebin.com/Xz4CRZpg 2011-12-21T20:20:10 i like that 2011-12-21T20:20:16 *** flowenol has quit IRC (Ping timeout: 240 seconds) 2011-12-21T20:20:21 skill, then mu, then high 2011-12-21T20:21:04 McLeopold: so help me I just had this "idea" where you could make a webpage where every time you visit you'd get the results of a db query... 2011-12-21T20:21:56 so, like, every web page in existance? 2011-12-21T20:22:02 right 2011-12-21T20:22:09 I think it might work 2011-12-21T20:22:55 btw, that top list is very stable 2011-12-21T20:23:04 for all three versions of "top" 2011-12-21T20:23:51 we would need to see quite a few upsets to start changing it, which is possible 2011-12-21T20:24:23 basically, if 2 bots' actual mu is a low and a high 2011-12-21T20:24:52 *** McLeopold has left #aichallenge 2011-12-21T20:25:36 I like how xathis has a high sigmas still too just because there's no one to bracket him 2011-12-21T20:26:30 it's lonely at hte top 2011-12-21T20:26:42 new goal, finish above momobot 2011-12-21T20:26:47 highest version 1 bot 2011-12-21T20:27:12 I think I got that one ok 2011-12-21T20:27:26 lot of people in the top 15 now who weren't in our idea of the top 10 2011-12-21T20:27:43 we're the only two version 1 bots in the top 100 atm 2011-12-21T20:27:48 yeah i know 2011-12-21T20:28:02 i'm curious as to whether that's just more time or better bots at the last minute or no idea? 2011-12-21T20:28:06 my v1 bot would be losing to starters as it timed out 2011-12-21T20:28:23 antimac: we haven't really had enough games up there yet 2011-12-21T20:28:28 *** gustavokambara has joined #aichallenge 2011-12-21T20:29:01 *** valydo has quit IRC (Ping timeout: 255 seconds) 2011-12-21T20:29:05 we only have 2 days left though 2011-12-21T20:29:19 well I'm hoping the game rate is going to pick up here 2011-12-21T20:29:26 there were hopes of 200 games at the top 2011-12-21T20:29:41 like runevision and komaki, they could well end up top 10 2011-12-21T20:29:59 that looks very unlikely at present 2011-12-21T20:30:01 any approx calcs how many games top 100 will have before the end 2011-12-21T20:30:32 there will be 36h left when the cutoff goes up next 2011-12-21T20:31:30 if they do another incremental slide up to the 500 point with 24h left 2011-12-21T20:32:02 top bots might hit 150-175 2011-12-21T20:32:10 10 minutes between games at 500 bots? 2011-12-21T20:32:20 am wondering if this time is enough to decide ranking. May be for ranks in 1000 I mean 2011-12-21T20:32:36 *** Garf has quit IRC (Quit: Make a new plan, Stan!) 2011-12-21T20:32:36 *** JorgeB has joined #aichallenge 2011-12-21T20:32:40 I'm not sure how precise you could ever be about 1000th place 2011-12-21T20:32:52 if you're in 1000 you may or may not be better 1v1 than 1001 and worse than 999 2011-12-21T20:32:57 so it's all statistical anyway 2011-12-21T20:33:12 true 2011-12-21T20:33:44 144 games in a day for the 500 2011-12-21T20:34:07 that would be good to see 2011-12-21T20:35:12 *** olexs has quit IRC (Quit: Leaving.) 2011-12-21T20:35:44 *** pairofdice has quit IRC (Quit: When cryptography is outlawed, bayl bhgynjf jvyy unir cevinpl.) 2011-12-21T20:35:46 *** creation has joined #aichallenge 2011-12-21T20:36:36 I feel bad to see my bot does very bad on 1v1 and normal mazes. It does well only in multihill mazes with close neighbor. It was a result of a last minute bug fix 2011-12-21T20:36:59 *** tobym has quit IRC (Remote host closed the connection) 2011-12-21T20:37:05 but if I didnt fix that, my bot wouldnt have made it to top 100 atleast for now 2011-12-21T20:40:48 *** creation has quit IRC (Quit: Page closed) 2011-12-21T20:41:57 *** creation has joined #aichallenge 2011-12-21T20:44:37 *** creation_ has joined #aichallenge 2011-12-21T20:45:39 *** creation has quit IRC (Client Quit) 2011-12-21T20:46:09 *** creation_ has quit IRC (Client Quit) 2011-12-21T20:46:40 yay, my bot survived a game without timeout! http://aichallenge.org/visualizer.php?game=300276&user=8660 2011-12-21T20:46:42 *** creation has joined #aichallenge 2011-12-21T20:57:52 *** gustavokambara has left #aichallenge 2011-12-21T20:59:35 i like the clock on the page, it makes things seem more exciting 2011-12-21T21:02:50 *** Zedenstein has quit IRC (Ping timeout: 258 seconds) 2011-12-21T21:12:17 I just noticed my blood pressure went up to 150/100 :S 2011-12-21T21:12:30 I just went on BP meds in time for this contest ;-) 2011-12-21T21:12:38 :P 2011-12-21T21:13:11 Am planning to leave for my native to spend some time with family and bring it under control 2011-12-21T21:15:57 definitly need a vacation 2011-12-21T21:18:28 BenJackson: meh, losing even more games :P 2011-12-21T21:18:38 I suck a _lot_ on small maps 2011-12-21T21:18:58 hedonism *does* have a negative connotation 2011-12-21T21:19:57 wow, I should have won that one against JG.WAS 2011-12-21T21:22:36 well, my bot sucked on small maps before that change 2011-12-21T21:25:25 contestbot: seen xathis 2011-12-21T21:25:25 amstan_: xathis was last seen in #aichallenge 5 hours, 24 minutes, and 2 seconds ago: mega1: yes, i try to have it ready when the finals end 2011-12-21T21:29:24 *** Westicles has joined #aichallenge 2011-12-21T21:32:17 *** tetrahedron has quit IRC (Quit: Page closed) 2011-12-21T21:37:44 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-12-21T21:40:28 *** xathis has quit IRC () 2011-12-21T21:41:02 *** creation has quit IRC (Quit: Ex-Chat) 2011-12-21T21:41:10 *** antimac has quit IRC (Ping timeout: 258 seconds) 2011-12-21T21:43:49 *** dmj111 has joined #aichallenge 2011-12-21T21:45:23 *** Westicles has quit IRC (Ping timeout: 258 seconds) 2011-12-21T21:45:24 *** analyst74 has joined #aichallenge 2011-12-21T21:47:18 *** xyxan has quit IRC (Ping timeout: 258 seconds) 2011-12-21T21:48:09 *** chris___0076 is now known as chris_0076 2011-12-21T21:49:30 damn, I should have taken screenshot when I was #4. :p 2011-12-21T21:50:57 *** FranPeruzzi has quit IRC (Quit: Page closed) 2011-12-21T21:53:21 what's a good alternative to dyndns? 2011-12-21T21:53:33 they apparently blocked my account, but they didn't even send an email or give a reason 2011-12-21T21:53:47 so my backups haven't been working all day, and I wouldn't have been able to ssh into my computer... 2011-12-21T21:54:02 I've used no-ip before 2011-12-21T21:54:32 thestinger: do you have your own domain? 2011-12-21T21:54:50 thestinger: i use dnsexit for that 2011-12-21T21:54:52 and it's free 2011-12-21T21:55:01 but you need to have a domain already 2011-12-21T21:55:04 amstan_: nope, I don't have a domain anymore 2011-12-21T21:55:16 I had one but I wasn't using it for anything other than ssh basically 2011-12-21T21:55:17 long term i recomend getting a domain 2011-12-21T21:55:25 short term, get noip 2011-12-21T21:56:02 thestinger, how did you enjoy being #1? 2011-12-21T21:56:12 analyst74: that was fun :) 2011-12-21T21:56:40 my bot has to climb back to ~25 now after a chain of mostly silly losses 2011-12-21T21:57:07 hopefully it will stop being an idiot xD 2011-12-21T21:57:12 good, I think you ended some where near 10th before reset right? 2011-12-21T21:57:18 *** FranPeruzzi has joined #aichallenge 2011-12-21T21:57:50 analyst74: well, I got to 11th with the upload I left up for around a week 2011-12-21T21:57:56 but that dropped back to ~15 2011-12-21T21:58:30 and then I didn't leave the other ones up long enough for the sigma to drop enough 2011-12-21T21:58:35 *** carpcrap has joined #aichallenge 2011-12-21T21:59:03 I got distracted in the final 2 weeks and didn't follow too closely, looking at the top 20 now, there are quite a few strange names 2011-12-21T21:59:46 *** Westicles has joined #aichallenge 2011-12-21T22:00:19 heh, BenJackson dropped down to 23 2011-12-21T22:00:24 I think there's a lot of churn going on atm 2011-12-21T22:00:34 I wish the old score wasn't reset, so I can tell who did last day submission and sky-rocketed to leader board 2011-12-21T22:01:36 ya, that's surprise too, hope it's just unlucky draws 2011-12-21T22:02:34 *** codetiger has quit IRC () 2011-12-21T22:08:29 *** analyst74 has quit IRC (Ping timeout: 240 seconds) 2011-12-21T22:09:15 *** analyst74 has joined #aichallenge 2011-12-21T22:13:23 *** TL_ has quit IRC (Quit: Page closed) 2011-12-21T22:20:27 aichallenge: Alexandru Stan epsilon * rb06cf24 / ants/visualizer/js/Application.js : 2011-12-21T22:20:27 aichallenge: Merge pull request #391 from avdg/fix/skipEventOnInputFocus 2011-12-21T22:20:27 aichallenge: Obfuscate document.activeElement in cases document doesn't exist (#390) - http://git.io/ijm9uQ 2011-12-21T22:21:46 *** TheLinker has joined #aichallenge 2011-12-21T22:24:57 flag is gaining 2011-12-21T22:25:03 irssi 2011-12-21T22:26:09 flagcapper? 2011-12-21T22:26:13 yes 2011-12-21T22:27:57 hmm, he lost ~1.7 mu in the past 7 hours 2011-12-21T22:28:37 gaining sigma, maybe :) 2011-12-21T22:31:47 http://tcpants.com/player/Accoun12301 play please 2011-12-21T22:32:04 @timeleft 2011-12-21T22:32:04 1 day, 22 hours, 27 minutes, and 56 seconds till winner is announced! 2011-12-21T22:32:23 @cutoff 2011-12-21T22:32:24 analyst74: No! 2011-12-21T22:32:36 @predict 2011-12-21T22:32:36 I predict raemde_ will win! 2011-12-21T22:32:49 analyst74: hmm, that could be cool 2011-12-21T22:32:53 amstan_: does the bot still play roulette? 2011-12-21T22:33:03 @roulette 2011-12-21T22:33:03 amstan_: *click* 2011-12-21T22:33:07 *** amstan_ is now known as amstan 2011-12-21T22:33:10 @roulette 2011-12-21T22:33:10 dmj111: *click* 2011-12-21T22:33:18 @roulette 2011-12-21T22:33:18 analyst74: *click* 2011-12-21T22:33:19 wow, I normally lost at that. 2011-12-21T22:33:27 dmj111: for some reason the services thing doesn't work, so he forgets to get himself op 2011-12-21T22:33:39 ah. 2011-12-21T22:34:28 analyst74: yeah, too complicated, no json for it 2011-12-21T22:34:29 who is raemde? 2011-12-21T22:34:31 analyst74: next contest 2011-12-21T22:34:33 @predict 2011-12-21T22:34:33 I predict Kingpin13 will win! 2011-12-21T22:34:41 ok :( 2011-12-21T22:35:21 oh, I thought predict actually do some predictive analysis. :p 2011-12-21T22:35:23 close enough: 2011-12-21T22:35:24 @cutoff 2011-12-21T22:35:25 amstan: cutoff could be http://aichallenge.org/server_stats.php. 2011-12-21T22:35:54 cool 2011-12-21T22:41:06 accoun: i am running a few games with my bot, antsinmypants 2011-12-21T22:41:59 going to bed though. i will run some more in the morning if you like. 2011-12-21T22:43:10 K 2011-12-21T22:43:26 this is funny, accoun does this every year, he keeps playing on tcp when the contest is over 2011-12-21T22:43:32 the force is strong in him, lol 2011-12-21T22:46:04 didn't he revert to starter on official? 2011-12-21T22:46:16 @help predict 2011-12-21T22:46:16 thestinger: (predict ) -- Alias for "echo I predict $randnick will win!". 2011-12-21T22:46:34 amstan: is that what you wanted $randnick for? :P 2011-12-21T22:46:39 for fun 2011-12-21T22:46:40 @predict xathis 2011-12-21T22:46:40 I predict Clex will win! 2011-12-21T22:56:35 *** ccsuser has joined #aichallenge 2011-12-21T22:57:10 *** JorgeB has quit IRC (Quit: Computer has gone to sleep.) 2011-12-21T22:58:07 russian don cocedate 2011-12-21T22:58:33 indeed 2011-12-21T22:58:40 in future canade will be rissian 2011-12-21T22:59:02 in future canada will be russian 2011-12-21T22:59:32 noooo~ I like Canada to be Chiliean 2011-12-21T22:59:46 not russian 2011-12-21T23:00:03 so we'll have winter all year long 2011-12-21T23:00:35 in houses well teperature 2011-12-21T23:00:44 and radiation protection 2011-12-21T23:01:16 55 sm walls 2011-12-21T23:03:16 actilay i have best bot 2011-12-21T23:03:39 actualay i have best bot 2011-12-21T23:04:35 ho is owner of http://tcpants.com ? 2011-12-21T23:07:08 who is owner of http://tcpants.com ? 2011-12-21T23:07:46 *** avdg has quit IRC (Quit: Leaving.) 2011-12-21T23:08:37 aiK0n play please, i have version for bet u in all maps 2011-12-21T23:11:59 *** avdg has joined #aichallenge 2011-12-21T23:13:15 a1k0n play please, i have version for bet u in all maps 2011-12-21T23:15:18 *** delineate has quit IRC (Quit: Page closed) 2011-12-21T23:15:31 *** delineate has joined #aichallenge 2011-12-21T23:15:33 *** FranPeruzzi has left #aichallenge 2011-12-21T23:15:53 i think it's a guy called T. C. Pants 2011-12-21T23:16:00 Accoun: You realize lots of people have published source? 2011-12-21T23:16:01 he named the website after himself :) 2011-12-21T23:16:04 you can run all the top bots yourself 2011-12-21T23:16:28 Accoun: https://github.com/a1k0n/ants 2011-12-21T23:16:57 my bot is the best 2011-12-21T23:17:09 *** u_ has quit IRC (Quit: u_) 2011-12-21T23:17:20 Accoun: opensource your bot and we'll see 2011-12-21T23:17:40 http://tcpants.com/player/Accoun20101 play with it 2011-12-21T23:18:20 *** avdg has joined #aichallenge 2011-12-21T23:30:44 turnlimit saved me 2011-12-21T23:30:55 *** teapotahedron_ni has quit IRC (Quit: Page closed) 2011-12-21T23:34:11 russian killer ants will own canada 2011-12-21T23:35:21 *** codetiger has joined #aichallenge 2011-12-21T23:35:24 *** ccsuser has quit IRC (Ping timeout: 258 seconds) 2011-12-21T23:35:31 Accoun: he's not even canadian.. 2011-12-21T23:35:58 *** UncleVasya has joined #aichallenge 2011-12-21T23:36:24 I'm getting a lot of games :-) either it likes me or there's a lot more servers online or they have cut a lot from the bottom 2011-12-21T23:36:53 Kettling: with 40 workers active.. 2011-12-21T23:36:57 Kettling: not surprising 2011-12-21T23:37:02 wow, two games in a row I would have benefitted from more turns 2011-12-21T23:37:43 I get a game every 45 mins 2011-12-21T23:37:58 until yesterday, I thought my bot was forgoten 2011-12-21T23:38:12 lol http://aichallenge.org/visualizer.php?game=311193&user=53 2011-12-21T23:38:50 UncleVasya: that ant is bored 2011-12-21T23:39:34 actually that's a very very reasonable play on that map 2011-12-21T23:39:50 there's a very reasonable course of action 2011-12-21T23:39:50 although I suppose both bots could back out of it to freedom 2011-12-21T23:39:55 Nice game... Seems those bots are too concerned about their territory 2011-12-21T23:40:03 you can squeeze bots out of that area just fine 2011-12-21T23:40:05 get some food 2011-12-21T23:40:08 then surround the enemy 2011-12-21T23:40:22 I should make a scenario and test my bot 2011-12-21T23:40:45 oh I can just pass it this input 2011-12-21T23:40:47 not really. Im my bot ants in battle aren't considered by another systems (food gathering, exploring etc) 2011-12-21T23:40:48 my bot would sacrifice under such conditions 2011-12-21T23:40:52 *In 2011-12-21T23:41:10 UncleVasya: you should be 2011-12-21T23:41:36 IS there any cut-off nearing? 2011-12-21T23:41:47 *** gaussin has joined #aichallenge 2011-12-21T23:41:56 idk, janzert's department 2011-12-21T23:41:58 @timeleft 2011-12-21T23:41:58 1 day, 21 hours, 18 minutes, and 2 seconds till winner is announced! 2011-12-21T23:42:13 oh, hm, just passing that input doesn't quite work 2011-12-21T23:42:20 *** gaussin has left #aichallenge 2011-12-21T23:42:21 BenJackson: why not? 2011-12-21T23:42:23 my bot would kill all the enemy having decided it was sessile 2011-12-21T23:42:32 sessile? 2011-12-21T23:42:35 not moving 2011-12-21T23:42:42 opposite of motile 2011-12-21T23:42:44 oh, lol 2011-12-21T23:43:38 BenJackson: i could put you in touch with that guy, see if you want to play against him on the same map 2011-12-21T23:43:56 I could just download the map and make a scenario 2011-12-21T23:44:03 but at this poitn I'm trying not to get caught up in bot hacking 2011-12-21T23:44:05 just let the finals go 2011-12-21T23:44:17 yeah, but you don't know his moves 2011-12-21T23:44:59 *** codetiger has quit IRC (Remote host closed the connection) 2011-12-21T23:45:11 maybe I'll play that map 2011-12-21T23:46:25 http://www.jmock.org/javadoc/2.5.1/org/hamcrest/Matcher.html#_dont_implement_Matcher___instead_extend_BaseMatcher_() 2011-12-21T23:47:48 @seen mcleopold 2011-12-21T23:47:48 BenJackson: mcleopold was last seen in #aichallenge 3 hours, 23 minutes, and 25 seconds ago: basically, if 2 bots' actual mu is a low and a high 2011-12-21T23:49:28 BenJackson: what's up? 2011-12-21T23:50:11 he's got a game count balance query 2011-12-21T23:51:55 * BenJackson pokes around the 3000 bots 2011-12-21T23:52:17 starters 2011-12-21T23:53:25 this person has 11 subs and moves less than a starter 2011-12-21T23:53:34 also username suffix '2' heh 2011-12-21T23:54:03 aha, there is a "real" kimriber too 2011-12-21T23:54:57 *** Guest74558 has quit IRC (Ping timeout: 258 seconds) 2011-12-21T23:55:30 I wonder if 2500s are tutorial bots? 2011-12-21T23:55:37 I don't know the tut bot behavior well enough 2011-12-21T23:55:45 but there are striking similarities to all these bots: 2011-12-21T23:55:54 http://aichallenge.org/visualizer.php?game=311851&user=12241 2011-12-21T23:56:55 oops that was 2000s even 2011-12-21T23:57:22 at least they don't seem to be killing themselves? 2011-12-21T23:57:33 I think that was part of the tutorial 2011-12-21T23:57:42 oh 2011-12-21T23:58:00 ok at 1500 people know hills at least sometimes 2011-12-21T23:59:05 sliding the cut to 2000 seems really safe