2011-12-08T00:00:43  *** delt0r_ <delt0r_!~bob@178-191-176-120.adsl.highway.telekom.at> has quit IRC (Ping timeout: 252 seconds)
2011-12-08T00:08:49  *** b0rder <b0rder!~b0rder@114.246.86.110> has quit IRC (Quit: 离开)
2011-12-08T00:09:09  *** besh <besh!835e7324@gateway/web/freenode/ip.131.94.115.36> has quit IRC (Quit: Page closed)
2011-12-08T00:13:31  *** delt0r_ <delt0r_!~bob@178-191-251-92.adsl.highway.telekom.at> has joined #aichallenge
2011-12-08T00:14:30  *** McLeopold <McLeopold!~Scott@c-98-247-248-39.hsd1.wa.comcast.net> has joined #aichallenge
2011-12-08T00:16:03  *** yoden <yoden!~horstman@cpe-76-190-250-126.neo.res.rr.com> has quit IRC (Quit: Leaving.)
2011-12-08T00:18:47  *** AntDroid_ <AntDroid_!781c39b6@gateway/web/freenode/ip.120.28.57.182> has joined #aichallenge
2011-12-08T00:20:20  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has quit IRC (Ping timeout: 252 seconds)
2011-12-08T00:24:52  *** avdg <avdg!~avdg@78-22-168-200.access.telenet.be> has joined #aichallenge
2011-12-08T00:25:19  *** Areks <Areks!~Areks@rs.gridnine.com> has joined #aichallenge
2011-12-08T00:29:30  *** replore_ <replore_!~replore@203.152.213.161.static.zoot.jp> has joined #aichallenge
2011-12-08T00:30:28  <greghaynes> Influence maps ftw!
2011-12-08T00:30:36  <McLeopold> \o/
2011-12-08T00:31:43  <greghaynes> Someone posted collaborative diffusion, but its really just an influence map
2011-12-08T00:32:49  <McLeopold> it seems like the second best method to me
2011-12-08T00:33:51  <greghaynes> Whats the first?
2011-12-08T00:34:09  <McLeopold> something that controls individual ants toward the best path
2011-12-08T00:34:31  <McLeopold> or, something where 2 ants coordinate together
2011-12-08T00:34:41  *** welterde <welterde!welterde@thinkbase.srv.welterde.de> has quit IRC (Ping timeout: 248 seconds)
2011-12-08T00:34:47  *** nickjohnson <nickjohnson!u789@gateway/web/irccloud.com/x-aiybqyxjhkugfabt> has quit IRC (Ping timeout: 240 seconds)
2011-12-08T00:34:51  <greghaynes> You can do that with most methods, thats basically the end goal
2011-12-08T00:34:51  *** Fluxid <Fluxid!~fluxid@schranz.io> has quit IRC (Ping timeout: 252 seconds)
2011-12-08T00:35:07  *** modafinil <modafinil!u515@gateway/web/irccloud.com/x-anxmulfgmqjstajz> has quit IRC (Ping timeout: 255 seconds)
2011-12-08T00:35:07  <greghaynes> I find influence mapping to coordinate much easier than per ant searching
2011-12-08T00:36:54  *** Fluxid <Fluxid!~fluxid@schranz.io> has joined #aichallenge
2011-12-08T00:36:59  <McLeopold> hmm, I think for exploring or finding chokepoints, they have their place, but not for effecient food gathering
2011-12-08T00:37:21  *** Fluxid is now known as Guest86987
2011-12-08T00:39:09  *** Jak_o_Shadows <Jak_o_Shadows!~hayden@CPE-144-136-214-52.sa.bigpond.net.au> has quit IRC (Remote host closed the connection)
2011-12-08T00:39:55  *** Jak_o_Shadows <Jak_o_Shadows!~hayden@CPE-144-136-214-52.sa.bigpond.net.au> has joined #aichallenge
2011-12-08T00:40:14  *** modafinil <modafinil!u515@gateway/web/irccloud.com/x-zfhsthufluhjbduu> has joined #aichallenge
2011-12-08T00:46:25  *** amstan <amstan!~amstan@69-196-130-230.dsl.teksavvy.com> has quit IRC (Quit: Konversation terminated!)
2011-12-08T00:46:28  *** amstan_ <amstan_!~amstan@69-196-130-230.dsl.teksavvy.com> has joined #aichallenge
2011-12-08T00:46:28  *** ChanServ sets mode: +o amstan_
2011-12-08T00:51:08  *** welterde <welterde!welterde@thinkbase.srv.welterde.de> has joined #aichallenge
2011-12-08T00:55:45  <greghaynes> Its pretty simple to make influene mapping generate equivilent data to a* or any of the similar searches
2011-12-08T00:56:11  <greghaynes> So not sure why its inefficient for food gathering
2011-12-08T00:56:50  <McLeopold> making sure 2 ants don't go for a pile when 1 ant would do
2011-12-08T00:57:23  <greghaynes> Yea, so you decrement your map based on occupancy
2011-12-08T00:57:30  <greghaynes> that is basically its strongpoint
2011-12-08T00:59:50  <greghaynes> The downside is that you dont get as quick response time for fast changing maps
2011-12-08T01:00:12  <greghaynes> but when the food and ant count grows it doesnt lose efficiency which is win
2011-12-08T01:00:42  <BenJackson> food count is never high
2011-12-08T01:01:14  <BenJackson> it never gets very high as an absolute value and it never gets very high relative to ant count after the beginning
2011-12-08T01:01:59  <greghaynes> ah, well the problem ive noticed with a*/etc is you are doing an exponential algo * num ants which doesnt work out well
2011-12-08T01:02:19  <BenJackson> if you search from food to ants it works fine
2011-12-08T01:02:25  <BenJackson> and food is important enough to handle specifically
2011-12-08T01:02:27  <greghaynes> sure, but then you have to collaborate
2011-12-08T01:02:28  <McLeopold> yeah, I'm working on bfs, search from food
2011-12-08T01:02:49  <BenJackson> well that *is* how i do it so I think it works ok :)
2011-12-08T01:02:56  <greghaynes> and food * num_ants is just as bad as num_ants * food ;)(
2011-12-08T01:03:22  <BenJackson> yes but you're searching from food to any ant
2011-12-08T01:03:24  <BenJackson> not every ant
2011-12-08T01:03:35  <McLeopold> BenJackson: how do you explorer? last seen? influece map?
2011-12-08T01:03:36  <antimatroid> any moveable ant*
2011-12-08T01:03:37  <BenJackson> and the distance is capped anyway
2011-12-08T01:03:39  <greghaynes> but you dont want the n closest ants?
2011-12-08T01:03:54  <BenJackson> McLeopold: various ways
2011-12-08T01:04:08  <BenJackson> primarily an influence-type map based on finding the unexplored frontier
2011-12-08T01:04:19  <McLeopold> do you detect dead ends?
2011-12-08T01:04:28  <antimatroid> i do iterative bfs from uncollected food to moveable ants, and atm i'm not collecting food that doesn't have a seen path to it from a hill
2011-12-08T01:04:34  <BenJackson> if you saw my multi-source bfs that was the basis of almost everything I did in my first decent version
2011-12-08T01:04:38  <antimatroid> well, i'm not searching over unseen land
2011-12-08T01:04:43  <BenJackson> you mean cul-de-sacs?
2011-12-08T01:05:13  <McLeopold> cul-de-sac implies a loop?
2011-12-08T01:05:38  <BenJackson> a path with no outlet
2011-12-08T01:05:52  <greghaynes> BenJackson: so you are doing an influence map?
2011-12-08T01:05:59  <BenJackson> in french literally "bott of bag"
2011-12-08T01:06:04  <BenJackson> +om
2011-12-08T01:06:19  <BenJackson> greghaynes: yes, but how I'm using it has varied considerably
2011-12-08T01:06:34  <BenJackson> originally I just did things like multi-source bfs sets of interesting things
2011-12-08T01:06:37  <BenJackson> food, enemy hills, etc
2011-12-08T01:06:47  <BenJackson> and then each ant went to the closest one of those
2011-12-08T01:06:50  <greghaynes> Makes sense, I couldnt see a better way to aggregate map data though
2011-12-08T01:06:59  <BenJackson> then that got modulated by scaling the closeness
2011-12-08T01:07:10  <BenJackson> eg so "hill defense" was only interesting very locally
2011-12-08T01:07:16  <BenJackson> while attacking was interesting far away
2011-12-08T01:07:34  <BenJackson> now I have those maps but I use them indirectly in various ways
2011-12-08T01:08:01  <BenJackson> since I can immediately ask "how many steps is this ant to *any* enemy ant"
2011-12-08T01:08:03  <BenJackson> not *every*
2011-12-08T01:13:18  <BenJackson> McLeopold: Anyway if you see a dead end then the frontier extends beyond the back wall of it and thus the influence for exploring no longer eminates from inside
2011-12-08T01:13:27  <BenJackson> in fact the gradient immediately flips around
2011-12-08T01:13:51  <BenJackson> exploring is not that hard
2011-12-08T01:13:57  <BenJackson> you just need some attractive force on unseen squares
2011-12-08T01:14:05  <BenJackson> what's harder is maintaining vision long term
2011-12-08T01:14:18  <BenJackson> that's a different problem from most in the game:
2011-12-08T01:14:23  <BenJackson> it's not "somewhere to go"
2011-12-08T01:23:49  *** ltriant <ltriant!~ltriant@lithium.mailguard.com.au> has quit IRC (Quit: Computer has gone to sleep)
2011-12-08T01:28:32  <BenJackson> once again brute force seems fast enough
2011-12-08T01:34:25  *** Scryer <Scryer!4a645367@gateway/web/freenode/ip.74.100.83.103> has quit IRC (Quit: Page closed)
2011-12-08T01:40:12  *** alc <alc!~nek@61.135.165.159> has joined #aichallenge
2011-12-08T01:46:03  *** JorgeB <JorgeB!~textual@c-98-234-188-88.hsd1.ca.comcast.net> has joined #aichallenge
2011-12-08T01:46:05  *** LouisMartin <LouisMartin!6017820a@gateway/web/freenode/ip.96.23.130.10> has joined #aichallenge
2011-12-08T01:55:54  *** bhasker <bhasker!~bhasker@99-162-151-97.lightspeed.sntcca.sbcglobal.net> has joined #aichallenge
2011-12-08T01:57:26  *** pairofdice <pairofdice!~pod@dsl-hkibrasgw2-ff67c300-50.dhcp.inet.fi> has joined #aichallenge
2011-12-08T01:57:58  *** Palmik <Palmik!~palmik@man-208-200.pks.muni.cz> has joined #aichallenge
2011-12-08T02:01:45  *** pairofdice <pairofdice!~pod@dsl-hkibrasgw2-ff67c300-50.dhcp.inet.fi> has quit IRC (Ping timeout: 252 seconds)
2011-12-08T02:01:54  *** alc <alc!~nek@61.135.165.159> has quit IRC (Remote host closed the connection)
2011-12-08T02:02:47  *** ildarkaa <ildarkaa!5b90be69@gateway/web/freenode/ip.91.144.190.105> has joined #aichallenge
2011-12-08T02:07:02  *** pairofdice <pairofdice!~pod@dsl-hkibrasgw2-ff67c300-50.dhcp.inet.fi> has joined #aichallenge
2011-12-08T02:07:09  *** AntDroid_ <AntDroid_!781c39b6@gateway/web/freenode/ip.120.28.57.182> has quit IRC (Ping timeout: 265 seconds)
2011-12-08T02:11:47  *** cuda <cuda!~pod@dsl-hkibrasgw2-ff67c300-50.dhcp.inet.fi> has joined #aichallenge
2011-12-08T02:13:19  *** Jak_o_Shadows <Jak_o_Shadows!~hayden@CPE-144-136-214-52.sa.bigpond.net.au> has quit IRC (Read error: Connection reset by peer)
2011-12-08T02:13:26  *** pairofdice <pairofdice!~pod@dsl-hkibrasgw2-ff67c300-50.dhcp.inet.fi> has quit IRC (Ping timeout: 248 seconds)
2011-12-08T02:16:34  *** cuda <cuda!~pod@dsl-hkibrasgw2-ff67c300-50.dhcp.inet.fi> has quit IRC (Ping timeout: 244 seconds)
2011-12-08T02:35:45  <BenJackson> http://paste.aichallenge.org/Xz0Gl/
2011-12-08T02:35:50  <BenJackson> don't know why I didn't just do that sooner
2011-12-08T02:37:56  *** bmh <bmh!ada4e181@fsf/member/bmh> has joined #aichallenge
2011-12-08T02:38:16  <BenJackson> bmh: I just pasted http://paste.aichallenge.org/Xz0Gl/
2011-12-08T02:38:39  <bmh> BenJackson: what am I looking for?
2011-12-08T02:38:45  <BenJackson> show orange's vis
2011-12-08T02:39:19  <bmh> it isn't showing anything?
2011-12-08T02:40:22  <BenJackson> that doesn't bring up a viewer?
2011-12-08T02:40:23  <BenJackson> works for me
2011-12-08T02:40:54  <bmh> It brings up a viewer, but showing the AI state doesn't do anything
2011-12-08T02:41:15  <bmh> it certainly doesn't like to kill hills
2011-12-08T02:41:45  <BenJackson> not ai state, just the fog of war
2011-12-08T02:42:10  <BenJackson> that bot isn't doing anything but optimizing fog of war with a minute interest in exploration
2011-12-08T02:42:24  <bmh> looks efficient :)
2011-12-08T02:42:58  <bmh> wow that looks efficient
2011-12-08T02:43:08  <bmh> BenJackson: can you run a bot on my tcp server? I'm waiting for an opponent
2011-12-08T02:43:21  <BenJackson> I think that's it
2011-12-08T02:43:30  <BenJackson> I set the font size for those windows so small I can't read them
2011-12-08T02:43:37  <bmh> thanks. I told it to play 15 games. Back in 10.
2011-12-08T02:43:40  <bmh> (minutes)
2011-12-08T02:48:44  *** darkf <darkf!~darkf_dix@unaffiliated/darkf> has quit IRC (Quit: Leaving)
2011-12-08T02:49:48  *** Conorach <Conorach!~cono@90.170.78.100> has joined #aichallenge
2011-12-08T02:52:32  <bmh> back -- BenJackson: my bot is a little... zergy.
2011-12-08T02:52:46  *** amstan__ <amstan__!~alex@69-196-130-230.dsl.teksavvy.com> has joined #aichallenge
2011-12-08T02:52:46  *** ChanServ sets mode: +o amstan__
2011-12-08T02:53:00  <BenJackson> sorry I quit running it earlier -- the tcpclient leaks some resoruces
2011-12-08T02:53:06  <BenJackson> so I was killing them all because I was having major other problems
2011-12-08T02:53:13  *** GeorgeJ <GeorgeJ!~GeorgeJ@86.124.203.104> has joined #aichallenge
2011-12-08T02:53:13  *** GeorgeJ <GeorgeJ!~GeorgeJ@unaffiliated/georgej> has joined #aichallenge
2011-12-08T02:57:08  <bmh> I am so psyched to have a bot that doesn't crash
2011-12-08T02:57:12  *** bmh <bmh!ada4e181@fsf/member/bmh> has quit IRC (Quit: sleep)
2011-12-08T03:01:32  *** amstan__ <amstan__!~alex@69-196-130-230.dsl.teksavvy.com> has quit IRC (Ping timeout: 252 seconds)
2011-12-08T03:01:55  <McLeopold> BenJackson: influence map?
2011-12-08T03:02:21  <BenJackson> only for the slight interest in exploring
2011-12-08T03:02:38  <BenJackson> ending up with vision of the whole map is just literally optimizing having vision of the whole map
2011-12-08T03:02:44  <BenJackson> which somehow I always assumed would be too expensive
2011-12-08T03:03:51  <McLeopold> hmm
2011-12-08T03:16:51  *** aerique <aerique!310225@xs8.xs4all.nl> has joined #aichallenge
2011-12-08T03:17:47  *** mikewintermute <mikewintermute!~mikewinte@188-223-213-21.zone14.bethere.co.uk> has joined #aichallenge
2011-12-08T03:18:10  *** mj41 <mj41!~mj41_nby@195.39.106.218> has joined #aichallenge
2011-12-08T03:18:58  *** Thinq <Thinq!~qq@84-253-243-204.bb.dnainternet.fi> has quit IRC (Ping timeout: 244 seconds)
2011-12-08T03:33:32  *** Thinq <Thinq!~qq@84-253-243-204.bb.dnainternet.fi> has joined #aichallenge
2011-12-08T03:38:13  *** epicmonkey <epicmonkey!~epicmonke@178.130.5.103> has joined #aichallenge
2011-12-08T03:45:32  *** praveen_ <praveen_!cf2e371d@gateway/web/freenode/ip.207.46.55.29> has joined #aichallenge
2011-12-08T03:52:38  *** Fandekasp <Fandekasp!~Fandekasp@p26153-ipngn100102yosemiya.okinawa.ocn.ne.jp> has quit IRC (Ping timeout: 248 seconds)
2011-12-08T03:56:42  *** jon1 <jon1!~Adium@2001:67c:90:764:fa1e:dfff:fedf:5943> has joined #aichallenge
2011-12-08T04:06:23  *** team_9 <team_9!a869085c@gateway/web/freenode/ip.168.105.8.92> has joined #aichallenge
2011-12-08T04:06:27  *** olexs <olexs!~Leviathan@drms-4d01513d.pool.mediaWays.net> has joined #aichallenge
2011-12-08T04:10:19  *** Guest86987 is now known as Fluxid
2011-12-08T04:11:27  *** team_9 <team_9!a869085c@gateway/web/freenode/ip.168.105.8.92> has quit IRC (Quit: Page closed)
2011-12-08T04:18:25  *** ikaros <ikaros!~ikaros@studpool-wlan-74-83.fs.fbi.h-da.de> has joined #aichallenge
2011-12-08T04:22:41  *** pairofdice <pairofdice!~pod@dsl-hkibrasgw2-ff67c300-50.dhcp.inet.fi> has joined #aichallenge
2011-12-08T04:32:21  *** FatboyPunk <FatboyPunk!d57d47b2@gateway/web/freenode/ip.213.125.71.178> has joined #aichallenge
2011-12-08T04:35:07  *** antimatroid1 <antimatroid1!~nick@098.014.dsl.vic.iprimus.net.au> has joined #aichallenge
2011-12-08T04:35:07  *** antimatroid <antimatroid!~nick@098.014.dsl.vic.iprimus.net.au> has quit IRC (Read error: Connection reset by peer)
2011-12-08T04:37:15  *** olexs <olexs!~Leviathan@drms-4d01513d.pool.mediaWays.net> has quit IRC (Quit: Leaving.)
2011-12-08T04:44:21  *** mviel_ <mviel_!~mviel@213.163.164.204> has joined #aichallenge
2011-12-08T04:48:01  *** mviel__ <mviel__!~mviel@213.163.164.204> has quit IRC (Ping timeout: 255 seconds)
2011-12-08T04:49:10  *** retybok <retybok!~retybok@gw.theralys.com> has quit IRC (Quit: leaving)
2011-12-08T05:00:27  *** replore_ <replore_!~replore@203.152.213.161.static.zoot.jp> has quit IRC (Remote host closed the connection)
2011-12-08T05:05:01  *** praveen_ <praveen_!cf2e371d@gateway/web/freenode/ip.207.46.55.29> has quit IRC (Ping timeout: 265 seconds)
2011-12-08T05:18:04  *** Areks <Areks!~Areks@rs.gridnine.com> has quit IRC (Ping timeout: 268 seconds)
2011-12-08T05:30:09  *** Areks <Areks!~Areks@rs.gridnine.com> has joined #aichallenge
2011-12-08T05:45:59  *** antimatroid1 <antimatroid1!~nick@098.014.dsl.vic.iprimus.net.au> has quit IRC (Ping timeout: 252 seconds)
2011-12-08T05:46:44  *** antimatroid <antimatroid!~nick@098.014.dsl.vic.iprimus.net.au> has joined #aichallenge
2011-12-08T05:49:34  *** Mystret <Mystret!~Mystret@188.142.37.201> has joined #aichallenge
2011-12-08T05:52:25  *** antimatroid <antimatroid!~nick@098.014.dsl.vic.iprimus.net.au> has quit IRC (Ping timeout: 244 seconds)
2011-12-08T05:53:05  *** antimatroid <antimatroid!~nick@098.014.dsl.vic.iprimus.net.au> has joined #aichallenge
2011-12-08T06:04:49  *** ikaros <ikaros!~ikaros@studpool-wlan-74-83.fs.fbi.h-da.de> has quit IRC (Ping timeout: 244 seconds)
2011-12-08T06:08:01  *** antimatroid <antimatroid!~nick@098.014.dsl.vic.iprimus.net.au> has quit IRC (Ping timeout: 252 seconds)
2011-12-08T06:14:19  *** antimatroid <antimatroid!~nick@141.114-27-211.dynamic.dsl.mel.iprimus.net.au> has joined #aichallenge
2011-12-08T06:19:24  *** antimatroid <antimatroid!~nick@141.114-27-211.dynamic.dsl.mel.iprimus.net.au> has quit IRC (Ping timeout: 244 seconds)
2011-12-08T06:19:57  *** antimatroid <antimatroid!~nick@141.114-27-211.dynamic.dsl.mel.iprimus.net.au> has joined #aichallenge
2011-12-08T06:21:07  *** mviel_ <mviel_!~mviel@213.163.164.204> has quit IRC (Ping timeout: 240 seconds)
2011-12-08T06:21:21  *** mviel_ <mviel_!~mviel@213.163.164.204> has joined #aichallenge
2011-12-08T06:22:24  *** mviel__ <mviel__!~mviel@236.227.73.86.rev.sfr.net> has joined #aichallenge
2011-12-08T06:24:12  *** mviel_ <mviel_!~mviel@213.163.164.204> has quit IRC (Read error: Operation timed out)
2011-12-08T06:24:17  *** imp5imp5 <imp5imp5!5e9f065a@gateway/web/freenode/ip.94.159.6.90> has quit IRC (Ping timeout: 265 seconds)
2011-12-08T06:31:35  *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge
2011-12-08T06:36:43  *** Palmik <Palmik!~palmik@man-208-200.pks.muni.cz> has quit IRC (Ping timeout: 252 seconds)
2011-12-08T06:37:46  *** UncleVasya <UncleVasya!~UncleVasy@46-133-242-153.dialup.umc.net.ua> has joined #aichallenge
2011-12-08T06:44:47  *** epicmonkey <epicmonkey!~epicmonke@178.130.5.103> has quit IRC (Ping timeout: 240 seconds)
2011-12-08T06:46:09  *** Palmik <Palmik!~palmik@man-208-200.pks.muni.cz> has joined #aichallenge
2011-12-08T07:02:36  *** Palmik <Palmik!~palmik@man-208-200.pks.muni.cz> has quit IRC (Ping timeout: 240 seconds)
2011-12-08T07:08:48  *** g0llum <g0llum!~g0llum@82.113.99.35> has joined #aichallenge
2011-12-08T07:22:58  *** ikaros <ikaros!~ikaros@studpool-wlan-74-83.fs.fbi.h-da.de> has joined #aichallenge
2011-12-08T07:23:45  *** nfn <nfn!~nfn@vs0205.flosoft-servers.net> has joined #aichallenge
2011-12-08T07:36:27  *** amstan_ <amstan_!~amstan@69-196-130-230.dsl.teksavvy.com> has quit IRC (Ping timeout: 240 seconds)
2011-12-08T07:40:24  *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal)
2011-12-08T07:45:43  *** Anilm3 <Anilm3!~anil@4.Red-79-146-241.dynamicIP.rima-tde.net> has joined #aichallenge
2011-12-08T07:51:11  *** mviel_ <mviel_!~mviel@213.163.164.204> has joined #aichallenge
2011-12-08T07:51:30  *** zoli1 <zoli1!~zoli@catv-80-98-34-34.catv.broadband.hu> has joined #aichallenge
2011-12-08T07:55:01  *** mviel__ <mviel__!~mviel@236.227.73.86.rev.sfr.net> has quit IRC (Ping timeout: 252 seconds)
2011-12-08T08:00:17  *** ildarkaa <ildarkaa!5b90be69@gateway/web/freenode/ip.91.144.190.105> has quit IRC (Quit: Page closed)
2011-12-08T08:00:56  *** delt0r_ <delt0r_!~bob@178-191-251-92.adsl.highway.telekom.at> has quit IRC (Ping timeout: 240 seconds)
2011-12-08T08:08:26  *** mviel_ is now known as vielmath
2011-12-08T08:08:42  *** vielmath is now known as mviel_
2011-12-08T08:10:33  *** Surya <Surya!58b273ad@gateway/web/freenode/ip.88.178.115.173> has joined #aichallenge
2011-12-08T08:12:17  *** Palmik <Palmik!~palmik@man-208-200.pks.muni.cz> has joined #aichallenge
2011-12-08T08:14:57  *** delt0r_ <delt0r_!~bob@178-190-41-115.adsl.highway.telekom.at> has joined #aichallenge
2011-12-08T08:19:14  *** thestinger <thestinger!~strcat@CPE98fc1165cb6f-CM00159a09b2d0.cpe.net.cable.rogers.com> has quit IRC (Quit: WeeChat 0.3.6)
2011-12-08T08:32:42  *** imp5imp5 <imp5imp5!5e9f065a@gateway/web/freenode/ip.94.159.6.90> has joined #aichallenge
2011-12-08T08:38:27  *** jeff_ <jeff_!75fdb999@gateway/web/freenode/ip.117.253.185.153> has joined #aichallenge
2011-12-08T08:38:53  *** jeff_ is now known as Guest78989
2011-12-08T08:42:58  <Guest78989> contestbot: hi
2011-12-08T08:42:59  <contestbot> Guest78989: User error, it's not my fault.
2011-12-08T08:43:32  <Guest78989> contestbot: ranking 10
2011-12-08T08:43:33  <contestbot> Guest78989: I'm sorry Dave, err Guest78989; I cannot 'ranking'.
2011-12-08T08:43:47  <Guest78989> contestbot: rankings 10
2011-12-08T08:43:48  <contestbot> Guest78989: Top 10 players: pguillory(92.6), xathis(91.1), ChrisH(89.6), ThisIsNotABug(87.7), lazarant(87.6), delineate(87.5), RVeerdonk(87.0), GreenTea(86.9), Migi32(86.2), BenJackson(85.9)
2011-12-08T08:43:50  *** TheLinker <TheLinker!~TheLinker@host16.190-231-108.telecom.net.ar> has joined #aichallenge
2011-12-08T08:43:59  <Guest78989> contestbot: bye
2011-12-08T08:44:00  <contestbot> Guest78989: I do not know about 'bye', but I do know about these similar topics: 'beta'
2011-12-08T08:44:10  *** Guest78989 <Guest78989!75fdb999@gateway/web/freenode/ip.117.253.185.153> has quit IRC (Quit: Page closed)
2011-12-08T08:49:22  <pairofdice> ...derp
2011-12-08T08:49:32  <pairofdice> @rankdings
2011-12-08T08:49:33  <contestbot> pairofdice: I'm sorry Dave, err pairofdice; I cannot 'rankdings'.
2011-12-08T08:49:36  <pairofdice> @rankings
2011-12-08T08:49:37  <contestbot> pairofdice: Top 10 players: pguillory(92.6), xathis(91.1), ChrisH(89.6), ThisIsNotABug(87.7), lazarant(87.6), delineate(87.5), RVeerdonk(87.0), GreenTea(86.9), Migi32(86.2), BenJackson(85.9)
2011-12-08T08:52:13  *** TheLinker <TheLinker!~TheLinker@host16.190-231-108.telecom.net.ar> has quit IRC (Quit: Bye)
2011-12-08T08:57:12  *** keithroe <keithroe!43a1fc8b@gateway/web/freenode/ip.67.161.252.139> has joined #aichallenge
2011-12-08T08:59:29  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has joined #aichallenge
2011-12-08T09:04:34  *** amstan_ <amstan_!~amstan@69-196-130-230.dsl.teksavvy.com> has joined #aichallenge
2011-12-08T09:04:35  *** ChanServ sets mode: +o amstan_
2011-12-08T09:11:02  *** imp5imp5 <imp5imp5!5e9f065a@gateway/web/freenode/ip.94.159.6.90> has quit IRC (Ping timeout: 265 seconds)
2011-12-08T09:11:47  *** Areks <Areks!~Areks@rs.gridnine.com> has quit IRC (Ping timeout: 268 seconds)
2011-12-08T09:14:54  *** FatboyPunk <FatboyPunk!d57d47b2@gateway/web/freenode/ip.213.125.71.178> has quit IRC (Ping timeout: 265 seconds)
2011-12-08T09:20:07  *** goffrie <goffrie!~goffrie@206-248-186-123.dsl.teksavvy.com> has quit IRC (Ping timeout: 244 seconds)
2011-12-08T09:20:36  *** goffrie <goffrie!~goffrie@69-196-176-93.dsl.teksavvy.com> has joined #aichallenge
2011-12-08T09:22:49  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has quit IRC (Ping timeout: 252 seconds)
2011-12-08T09:29:17  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has joined #aichallenge
2011-12-08T09:30:48  *** avdg <avdg!~avdg@78-22-168-200.access.telenet.be> has quit IRC (Quit: Leaving.)
2011-12-08T09:32:54  *** tobym <tobym!~tobym@pool-71-183-248-125.nycmny.fios.verizon.net> has joined #aichallenge
2011-12-08T09:34:03  *** modafinil is now known as BEASTMODE
2011-12-08T09:34:28  *** sigh <sigh!~sigh@ppp114-153.static.internode.on.net> has quit IRC (Remote host closed the connection)
2011-12-08T09:38:06  *** keithroe <keithroe!43a1fc8b@gateway/web/freenode/ip.67.161.252.139> has quit IRC (Ping timeout: 265 seconds)
2011-12-08T09:38:15  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has quit IRC (Ping timeout: 252 seconds)
2011-12-08T09:51:14  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has joined #aichallenge
2011-12-08T09:54:48  *** thestinger <thestinger!~strcat@CPE98fc1165cb6f-CM00159a09b2d0.cpe.net.cable.rogers.com> has joined #aichallenge
2011-12-08T10:00:08  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has quit IRC (Ping timeout: 240 seconds)
2011-12-08T10:05:31  *** olexs <olexs!~Adium@s0527.dyn.hrz.tu-darmstadt.de> has joined #aichallenge
2011-12-08T10:05:52  <olexs> what's up with aichallenge workers? game rate is near zero since yesterday...
2011-12-08T10:08:32  *** NightExcessive <NightExcessive!~nightexce@unaffiliated/nightexcessive> has joined #aichallenge
2011-12-08T10:08:42  <Minthos> they scale up and down according to price fluctuations in the cloud
2011-12-08T10:12:42  <rwest> http://aichallenge.org/server_stats.php
2011-12-08T10:12:47  <rwest> only one server currently
2011-12-08T10:14:09  *** ikaros <ikaros!~ikaros@studpool-wlan-74-83.fs.fbi.h-da.de> has quit IRC (Quit: Ex-Chat)
2011-12-08T10:19:08  <olexs> not very nice scaling the workers down now this much, when the challenge is nearly over and people want to see current ranking of their latest bots before final submissions...
2011-12-08T10:19:57  <amstan_> olexs: when this happens price is usually 20 times more than normal
2011-12-08T10:20:09  <amstan_> it's really stupid to be playing games at that price
2011-12-08T10:20:22  <pairofdice> 20 times the price isn't nice either
2011-12-08T10:20:25  *** olexs <olexs!~Adium@s0527.dyn.hrz.tu-darmstadt.de> has quit IRC (Quit: Leaving.)
2011-12-08T10:20:37  <amstan_> yeah, i think it's like 5c/hour on average
2011-12-08T10:20:43  <amstan_> our limit is $1
2011-12-08T10:21:15  <rwest> amstan_: will the final tournament have more servers for more games in that week?
2011-12-08T10:21:33  <amstan_> yes, we conserved money to play much more games for the finals
2011-12-08T10:21:41  <pairofdice> Easier would be to cut away the low mu low sigma bots
2011-12-08T10:21:49  <amstan_> pairofdice: that too
2011-12-08T10:26:11  *** olexs <olexs!~Adium@s0823.dyn.hrz.tu-darmstadt.de> has joined #aichallenge
2011-12-08T10:28:54  *** jon1 <jon1!~Adium@2001:67c:90:764:fa1e:dfff:fedf:5943> has quit IRC (Quit: Leaving.)
2011-12-08T10:33:10  *** Karlis <Karlis!d5af6a0f@gateway/web/freenode/ip.213.175.106.15> has joined #aichallenge
2011-12-08T10:36:20  *** AntDroid <AntDroid!70c6400f@gateway/web/freenode/ip.112.198.64.15> has quit IRC (Quit: Page closed)
2011-12-08T10:37:56  *** NightExcessive <NightExcessive!~nightexce@unaffiliated/nightexcessive> has quit IRC (Quit: This computer has gone to sleep)
2011-12-08T10:45:11  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has joined #aichallenge
2011-12-08T10:46:31  <olexs> wow, the prices fluctuate so much? then I remove my inquiry :)
2011-12-08T10:46:42  *** mviel__ <mviel__!~mviel@236.227.73.86.rev.sfr.net> has joined #aichallenge
2011-12-08T10:46:52  <olexs> I thought it's a few percent more or less
2011-12-08T10:47:09  <rwest> google should sponser by giving servers!
2011-12-08T10:47:25  *** aerique <aerique!310225@xs8.xs4all.nl> has quit IRC (Quit: ...)
2011-12-08T10:47:26  <amstan_> rwest: that wouldn't be as easy as amazon
2011-12-08T10:48:32  <rwest> doesn't google have some scaling cloud service? I coulda swore I read about it
2011-12-08T10:49:50  <amstan_> rwest: yes, but it was probably not intended for outside use
2011-12-08T10:50:08  *** mviel_ <mviel_!~mviel@213.163.164.204> has quit IRC (Ping timeout: 244 seconds)
2011-12-08T10:50:17  <Minthos> they have something called Google Apps which is public
2011-12-08T10:50:20  <rwest> Well it looks like it could be, but you couldn't compile and the like on the server
2011-12-08T10:50:31  <amstan_> Minthos: that's only for python and go
2011-12-08T10:50:35  <rwest> it seems it's web apps only
2011-12-08T10:50:40  <rwest> yea that's what I just read
2011-12-08T10:51:15  <Minthos> so unless you could virtualize a worker in a python or go process you're screwed :p
2011-12-08T10:51:19  <amstan_> we need real linux servers for all the 25 languages we have
2011-12-08T10:51:31  *** AntDroid <AntDroid!70c6400f@gateway/web/freenode/ip.112.198.64.15> has joined #aichallenge
2011-12-08T10:51:35  <amstan_> and main server frontend is in php
2011-12-08T10:51:40  <amstan_> and even if it was in python
2011-12-08T10:51:54  <amstan_> last time i checked google apps only liked python 2.5
2011-12-08T10:52:08  <amstan_> which is like 3 years old
2011-12-08T10:53:18  <rwest> it's nuts that amazon prices fluctuate so much
2011-12-08T10:53:50  <amstan_> if you think about it, it's supercheap
2011-12-08T10:53:59  <Minthos> seems reasonable to me
2011-12-08T10:54:12  <amstan_> we spent 2000 bucks for like 7 servers running at the same time on average, with 100% cpu usage
2011-12-08T10:54:28  <Surya> Thx a lot for this awesome challenge anyway amstan :] I enjoy it too much (literally, my professionnal productivity & personal life get fucked up :])
2011-12-08T10:54:42  <Minthos> amstan_: 2k/7 servers for how long?
2011-12-08T10:54:49  <amstan_> the length of the contest
2011-12-08T10:55:00  <Minthos> two-three months?
2011-12-08T10:55:05  <amstan_> yeah
2011-12-08T10:55:30  <Minthos> I think 7 regular vpses would have been cheaper
2011-12-08T10:55:37  <Minthos> with similar resources
2011-12-08T10:55:52  <amstan_> similar resources? 17GB of ram each and dual core xeons?
2011-12-08T10:55:59  <Minthos> hmm.. perhaps not
2011-12-08T10:56:05  <Minthos> but the memory seems overkill
2011-12-08T10:56:09  <amstan_> and an interface to start and stop them at will?
2011-12-08T10:56:12  <amstan_> not really
2011-12-08T10:56:18  <amstan_> you have 1.5B limit per bot
2011-12-08T10:56:27  <amstan_> you have at most 8 bots running at the same time
2011-12-08T10:56:34  <amstan_> 1.5GB*
2011-12-08T10:56:46  <Minthos> yeah 1.5 GB per bot seems overkill
2011-12-08T10:56:47  <amstan_> so that's 12GB
2011-12-08T10:56:55  <amstan_> Minthos: java doesn't start otherwise
2011-12-08T10:57:13  *** dvladim <dvladim!~dvladim@85.113.131.65> has joined #aichallenge
2011-12-08T10:57:14  <Minthos> huh?
2011-12-08T10:57:33  <rwest> I think being lent 1 of google's server for the contest would have been better ^^
2011-12-08T10:57:47  <rwest> some quad octeon or better with 64 cores
2011-12-08T10:57:49  <Minthos> java doesn't require 1.5 GB
2011-12-08T10:58:01  <amstan_> Minthos: well, idk, ask janzert for more details
2011-12-08T10:58:43  *** ikaros <ikaros!~ikaros@dslb-084-059-073-057.pools.arcor-ip.net> has joined #aichallenge
2011-12-08T10:58:45  <amstan_> also.. being able to bring up or down the servers at will and reverting to the blank ubuntu to reinstall the worker is a very nice feature
2011-12-08T10:58:46  <Minthos> ok but anyway, it's not unreasonably priced
2011-12-08T10:58:56  *** besh <besh!835e7324@gateway/web/freenode/ip.131.94.115.36> has joined #aichallenge
2011-12-08T10:59:10  <Minthos> and cloud computing is popular for people who have workloads that flucutate a lot
2011-12-08T10:59:23  <Minthos> so amazon needs to have capacity to handle peak demand
2011-12-08T10:59:41  <Minthos> so they price resources higher when there is much demand, and lower when there is less
2011-12-08T10:59:57  <Minthos> makes perfect sense :)
2011-12-08T11:00:09  <pairofdice> Pretty much
2011-12-08T11:00:36  *** avdg <avdg!~avdg@78-22-168-200.access.telenet.be> has joined #aichallenge
2011-12-08T11:01:55  <Minthos> my bot uses 24 megs on a 120x120 map
2011-12-08T11:02:23  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has quit IRC (Ping timeout: 252 seconds)
2011-12-08T11:02:49  <rwest> Minthos: that's about what mine uses
2011-12-08T11:03:12  <rwest> well actually
2011-12-08T11:03:16  <rwest> less
2011-12-08T11:04:08  *** Surya <Surya!58b273ad@gateway/web/freenode/ip.88.178.115.173> has quit IRC (Ping timeout: 265 seconds)
2011-12-08T11:06:04  *** decaz <decaz!4a71ee12@gateway/web/freenode/ip.74.113.238.18> has quit IRC (Ping timeout: 265 seconds)
2011-12-08T11:11:30  *** mikewintermute <mikewintermute!~mikewinte@188-223-213-21.zone14.bethere.co.uk> has quit IRC (Quit: mikewintermute)
2011-12-08T11:12:12  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has joined #aichallenge
2011-12-08T11:16:42  <besh> can't you move to a food location ? I got a move blocked error
2011-12-08T11:17:00  <pairofdice> You can't
2011-12-08T11:17:15  <thestinger> food spawns at the end of the turn, and you're told about it at the start of the turn
2011-12-08T11:17:19  <thestinger> so it can be next to your ants
2011-12-08T11:17:33  <rwest> don't you automatically collect them when near?
2011-12-08T11:17:40  <rwest> I have never had this problem and I should
2011-12-08T11:17:42  <thestinger> yes, during the turn - not in between them
2011-12-08T11:17:58  <besh> well i thought so too but food can be dropped close to your ant I think..
2011-12-08T11:18:01  <rwest> unless!
2011-12-08T11:18:14  <rwest> maybe the java start package calls i occupied
2011-12-08T11:18:16  <thestinger> food can spawn next to one of your ants at the end of the turn (when food spawns :P)
2011-12-08T11:18:18  <rwest> it
2011-12-08T11:18:32  <thestinger> and then it will be 1 distance from your ant at the start of the turn
2011-12-08T11:18:38  <thestinger> and movement happens before food is gathered
2011-12-08T11:18:42  <besh> thestinger: yes
2011-12-08T11:19:40  <rwest> hmm I have never seen the issue yet I have nothing preventing it
2011-12-08T11:20:00  <thestinger> it doesn't really matter, your move will just be blocked and you'll gather the food
2011-12-08T11:20:33  <rwest> well still a wast of a turn
2011-12-08T11:20:33  <thestinger> I wrote a special case to deal with it just to get rid of the warnings though
2011-12-08T11:20:42  <besh> yes that is what happened.
2011-12-08T11:20:58  <rwest> I can just modify the check to see if it's unoccupied
2011-12-08T11:21:01  <besh> thestinger: wow too cautions :)
2011-12-08T11:21:09  <besh> cautious*
2011-12-08T11:21:35  <thestinger> well, if 2 ants are next to a food I only want one to stay
2011-12-08T11:21:54  <thestinger> well I guess that would happen anyway if I ignored food blocking movement...
2011-12-08T11:22:48  <thestinger> I think I might improve my bot again today, for the first time in over a week :P
2011-12-08T11:23:28  <besh> good for u. mine seem to go down every week..
2011-12-08T11:23:51  <thestinger> just going to make combat _really_ per battle and keep generating moves until time runs out
2011-12-08T11:23:57  <thestinger> atm I never use more than 35ms
2011-12-08T11:24:31  <rwest> return ((this == LAND || this == DEAD) && this != FOOD);
2011-12-08T11:24:32  <rwest> there
2011-12-08T11:24:48  <rwest> actually
2011-12-08T11:24:48  <rwest> nvm
2011-12-08T11:24:51  <rwest> thats dumvb
2011-12-08T11:24:54  <rwest> it already did it
2011-12-08T11:25:35  * rwest realizes it was an enum and did it anyway
2011-12-08T11:25:54  <thestinger> which language are you using?
2011-12-08T11:26:00  <rwest> java
2011-12-08T11:26:05  *** QuirionPT <QuirionPT!~quirion@gyf5.kyla.fi> has joined #aichallenge
2011-12-08T11:26:15  <rwest> if it's labeled land it has no food on it
2011-12-08T11:26:19  <thestinger> ah
2011-12-08T11:26:31  <besh> yes it excludes that but my A* excludes only water. Got to fix that
2011-12-08T11:26:52  <thestinger> the food won't be there once gathered though
2011-12-08T11:27:08  <thestinger> so you might as well pathfind through it
2011-12-08T11:27:14  <rwest> add it to the isPassible() if you are using java besh
2011-12-08T11:27:26  <rwest> assuming you use that to check for water
2011-12-08T11:28:00  <besh> I assume food is passable for A*
2011-12-08T11:28:01  <McLeopold> I think the only time food would block a path is if it spawned next to an ant
2011-12-08T11:28:13  <thestinger> McLeopold: yeah :)
2011-12-08T11:28:15  <McLeopold> otherwise walking "through" it will gather it and not slow you down
2011-12-08T11:28:27  <rwest> I check my moves right before do them
2011-12-08T11:28:36  <thestinger> if I didn't handle food spawning next to ants, I think it would end up being the same behavior as what I do now
2011-12-08T11:28:40  <rwest> as a secondary check
2011-12-08T11:28:48  <thestinger> because I would try to have one of the ants it spawned next to move onto it
2011-12-08T11:28:50  <thestinger> and they would stay there
2011-12-08T11:29:05  <thestinger> which is exactly what I do now (I make one of the ants next to it stay there to gather it)
2011-12-08T11:29:59  <matthewd> thestinger: Though if you think they're moving, you'd risk a collision
2011-12-08T11:30:23  <thestinger> ah yeah, that's why I fixed it :)
2011-12-08T11:30:25  <rwest> if it happens, my ant will stay there or move a space along in it the direction of it's next target
2011-12-08T11:30:27  <thestinger> now I remember xD
2011-12-08T11:30:44  <rwest> it in
2011-12-08T11:38:03  *** olexs <olexs!~Adium@s0823.dyn.hrz.tu-darmstadt.de> has quit IRC (Quit: Leaving.)
2011-12-08T11:38:57  *** analyst74 <analyst74!~analyst74@96.53.97.142> has joined #aichallenge
2011-12-08T11:42:53  <besh> new combat strategy memetix using influence maps http://forums.aichallenge.org/viewtopic.php?f=24&t=2083
2011-12-08T11:48:54  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has quit IRC (Ping timeout: 248 seconds)
2011-12-08T11:50:38  *** dvladim <dvladim!~dvladim@85.113.131.65> has quit IRC (Ping timeout: 240 seconds)
2011-12-08T11:56:45  <besh> Anyone understood why he said the influence map should be computed after 1 move ? why not just compute it  from where the ants are now
2011-12-08T11:59:27  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has joined #aichallenge
2011-12-08T12:00:09  <pairofdice> Well, you don't want to know what happened but what will happen if I move there or here
2011-12-08T12:00:33  <besh> so is it attack_radius + 2 ?
2011-12-08T12:00:55  *** mleise <mleise!~marco@dslb-088-070-148-091.pools.arcor-ip.net> has joined #aichallenge
2011-12-08T12:01:11  <a1k0n> it's more like, all possible moves convolved with attack_radius
2011-12-08T12:01:24  <pairofdice> ^
2011-12-08T12:01:25  <a1k0n> you could precompute this in a 16-entry table
2011-12-08T12:02:18  <pairofdice> That's what I did, but didn't bother with water blocking movement or any other details either so it's... yeah
2011-12-08T12:03:08  *** SimonT <SimonT!~chatzilla@unaffiliated/simont> has joined #aichallenge
2011-12-08T12:03:29  <a1k0n> i figured this sort of approach would be smarter and better than what i did but i never worked out the details
2011-12-08T12:03:51  <pairofdice> It's simple and cheap
2011-12-08T12:03:58  <besh> yes that is what I do too.
2011-12-08T12:04:17  <besh> second questions: enemies = total[row][col] - influence[player][row][col]
2011-12-08T12:04:45  <besh> "We can now work out the best ant in each combat zone"
2011-12-08T12:04:49  *** tobel <tobel!~user@pD9E8A32A.dip.t-dialin.net> has joined #aichallenge
2011-12-08T12:04:59  <besh> best for what ?
2011-12-08T12:05:11  *** analyst74 <analyst74!~analyst74@96.53.97.142> has quit IRC (Quit:  Want to be different? Try HydraIRC -> http://www.hydrairc.com <-)
2011-12-08T12:05:20  *** Simon14 <Simon14!~chatzilla@unaffiliated/simont> has joined #aichallenge
2011-12-08T12:06:02  *** nplus <nplus!~chatzilla@unaffiliated/simont> has quit IRC (Ping timeout: 252 seconds)
2011-12-08T12:06:04  *** Simon14 is now known as nplus
2011-12-08T12:06:09  <pairofdice> " the best ant being the one that is fighting the least number of other ants"
2011-12-08T12:06:23  <pairofdice> Roght after your text
2011-12-08T12:06:27  <pairofdice> :p
2011-12-08T12:07:41  <besh> we are making coordinated moves of ant which doesn't require selection
2011-12-08T12:07:44  <besh> brb..
2011-12-08T12:08:08  *** bluegaspode <bluegaspode!55b223f9@gateway/web/freenode/ip.85.178.35.249> has joined #aichallenge
2011-12-08T12:08:36  *** SimonT <SimonT!~chatzilla@unaffiliated/simont> has quit IRC (Ping timeout: 240 seconds)
2011-12-08T12:10:43  *** olexs <olexs!~Leviathan@drms-4d01513d.pool.mediaWays.net> has joined #aichallenge
2011-12-08T12:14:31  <thestinger> I obviously overcomplicated combat...
2011-12-08T12:14:35  <thestinger> lemme try this way and compare
2011-12-08T12:15:56  *** Karlis <Karlis!d5af6a0f@gateway/web/freenode/ip.213.175.106.15> has quit IRC (Quit: Page closed)
2011-12-08T12:17:10  *** mj41 <mj41!~mj41_nby@195.39.106.218> has quit IRC (Ping timeout: 248 seconds)
2011-12-08T12:18:41  *** Hexren <Hexren!~hexren@46.59.196.52> has joined #aichallenge
2011-12-08T12:18:43  *** Hexren <Hexren!~hexren@46.59.196.52> has left #aichallenge
2011-12-08T12:19:44  *** Scryer <Scryer!4a645367@gateway/web/freenode/ip.74.100.83.103> has joined #aichallenge
2011-12-08T12:22:32  *** JorgeB <JorgeB!~textual@c-98-234-188-88.hsd1.ca.comcast.net> has quit IRC (Ping timeout: 252 seconds)
2011-12-08T12:23:18  *** RedTurtle <RedTurtle!48c24371@gateway/web/freenode/ip.72.194.67.113> has joined #aichallenge
2011-12-08T12:23:57  <rwest> my combat that I am working on is very similar to what memetix posted
2011-12-08T12:25:56  *** McLeopold <McLeopold!~Scott@c-98-247-248-39.hsd1.wa.comcast.net> has quit IRC (Quit: Leaving.)
2011-12-08T12:25:56  *** mviel__ <mviel__!~mviel@236.227.73.86.rev.sfr.net> has quit IRC (Read error: Connection timed out)
2011-12-08T12:26:25  *** kire <kire!~kire@78-22-125-201.access.telenet.be> has joined #aichallenge
2011-12-08T12:26:40  *** mviel__ <mviel__!~mviel@213.163.164.204> has joined #aichallenge
2011-12-08T12:30:57  *** dvladim <dvladim!~dvladim@85.113.131.65> has joined #aichallenge
2011-12-08T12:33:06  *** kire <kire!~kire@78-22-125-201.access.telenet.be> has quit IRC (Remote host closed the connection)
2011-12-08T12:34:43  *** GarfTop <GarfTop!~Garf@78-22-80-252.access.telenet.be> has joined #aichallenge
2011-12-08T12:34:54  <Scryer> @rwest Mine will be too, soon. :)  It fits better into my structure than a1k0n's.
2011-12-08T12:34:55  <contestbot> Scryer: An error occured while trying to show the previous error.
2011-12-08T12:35:06  *** GarfTop <GarfTop!~Garf@78-22-80-252.access.telenet.be> has quit IRC (Read error: Connection reset by peer)
2011-12-08T12:35:43  *** GarfTop <GarfTop!~Garf@78-22-80-252.access.telenet.be> has joined #aichallenge
2011-12-08T12:36:15  *** Garf <Garf!~Garf@78-22-80-252.access.telenet.be> has quit IRC (Ping timeout: 240 seconds)
2011-12-08T12:36:37  *** JorgeB <JorgeB!~textual@209.66.74.34> has joined #aichallenge
2011-12-08T12:36:56  <rwest> Scryer: I am using influence maps for my entire bot hah
2011-12-08T12:37:06  <thestinger> hm, I think memetix does things a bit differently than how I coded it
2011-12-08T12:37:50  <rwest> too... much... kfc...
2011-12-08T12:37:50  <rwest> ugh
2011-12-08T12:38:53  <thestinger> so for the "best ant", he just means ant within attackradius+1 with the minimum number of enemies in range, right?
2011-12-08T12:39:04  <thestinger> I already do that
2011-12-08T12:39:43  <thestinger> oh well, time to figure out why my bot isn't nearly that good :P
2011-12-08T12:42:00  <a1k0n> his method seems like it works well in mazes and confined areas but is too conservative in open spaces
2011-12-08T12:43:09  <BenJackson> I might implement it just to have it to fight against
2011-12-08T12:43:23  <thestinger> yeah, I'm implementing it right now just so I can compare
2011-12-08T12:43:37  <a1k0n> it's similar to what you're doing anyway Ben isn't it?  at last for enemy ants?
2011-12-08T12:43:45  <a1k0n> s/last/least/
2011-12-08T12:44:59  <BenJackson> mine is probably more similar to what I've heard peopel describe here where they assume all enemies are stationary
2011-12-08T12:45:10  <BenJackson> i just multiply all enemies by everywhere they could go
2011-12-08T12:45:13  <thestinger> ah
2011-12-08T12:45:16  <BenJackson> and fight against all those at once
2011-12-08T12:45:19  <thestinger> I just move them towards me (without collisions)
2011-12-08T12:45:29  <thestinger> and pretend they are staying still (so 2 scenarios)
2011-12-08T12:45:30  <a1k0n> that's pretty much what the influence map is
2011-12-08T12:45:41  <BenJackson> yes, I agree
2011-12-08T12:46:02  <BenJackson> but then I consider individual sets of moves against that
2011-12-08T12:46:08  <thestinger> this influence map method is a lot more cautious than mine
2011-12-08T12:46:20  <BenJackson> if my code lacks anything it's probabilities for where the enemies would go
2011-12-08T12:46:21  <thestinger> because it basically multiplies ants in all directions
2011-12-08T12:46:29  <BenJackson> so I might implement his combat to weight the probabilities of the enemy ants
2011-12-08T12:46:49  <a1k0n> the thing is, it's equally cautious and reckless cause he multiplies his own ants the same way
2011-12-08T12:46:56  <rwest> lol implement his combat so you can predict his moves
2011-12-08T12:47:01  <BenJackson> http://paste.aichallenge.org/Xz0Gl/
2011-12-08T12:47:07  <BenJackson> rwest: well if he can do it in a few ms in java...
2011-12-08T12:47:26  <BenJackson> btw that's what I finally did last night
2011-12-08T12:47:41  <BenJackson> I was inspired by the fog of war at the end of a pguillory game to just go ahead and directly maximize vis
2011-12-08T12:47:53  <BenJackson> so if you slide to the end there and hit fog you can see how nicely it works out
2011-12-08T12:47:56  <rwest> I have a similar combat on my bot and I couldn't figure out why it sucked
2011-12-08T12:48:03  <rwest> I need to separate it out from my main map
2011-12-08T12:48:07  <BenJackson> didn't cost any more than my other method which was far more dense
2011-12-08T12:48:12  <a1k0n> that's pretty much what i've been doing all along
2011-12-08T12:48:14  <rwest> since I spread the values too much
2011-12-08T12:48:24  <a1k0n> my ants don't bother scouting over known water either
2011-12-08T12:48:56  <BenJackson> my currently submitted bot tried to be within some number of steps (4? 5? I forget) of everything
2011-12-08T12:49:03  <BenJackson> it works well for some things
2011-12-08T12:49:10  <BenJackson> but it's too dense far from teh fight
2011-12-08T12:49:15  *** thestinger <thestinger!~strcat@CPE98fc1165cb6f-CM00159a09b2d0.cpe.net.cable.rogers.com> has quit IRC (Read error: Operation timed out)
2011-12-08T12:49:32  <a1k0n> yeah it's a tradeoff.  just getting full map coverage with a minimum number of ants works really well, except your ants are extremely vulnerable
2011-12-08T12:49:56  <pairofdice> If you have ants moving from hills you don't have to have stationary ants at all, if you keep track of the perimeter
2011-12-08T12:50:04  *** ademar_ <ademar_!~ademar@174.117.80.15> has quit IRC (Read error: Connection reset by peer)
2011-12-08T12:50:18  <BenJackson> I only worked out yesterday that it really takes 4v1 with ideal placement to beat a lone ant "in space"
2011-12-08T12:50:35  *** Garf <Garf!~Garf@78-22-80-252.access.telenet.be> has joined #aichallenge
2011-12-08T12:50:35  *** GarfTop <GarfTop!~Garf@78-22-80-252.access.telenet.be> has quit IRC (Ping timeout: 240 seconds)
2011-12-08T12:50:56  *** ademar_ <ademar_!~ademar@CPE0026f3379398-CM0026f3379395.cpe.net.cable.rogers.com> has joined #aichallenge
2011-12-08T12:50:58  <pairofdice> Yeah, in unbostructed map
2011-12-08T12:50:59  <a1k0n> not 3?
2011-12-08T12:51:19  <BenJackson> you can force the lone ant to die with 3 but it can trade you
2011-12-08T12:51:59  <BenJackson> anyway my point being that my vision of my dense web of ants being like white blood cells didn't pan out
2011-12-08T12:52:03  *** thestinger <thestinger!~strcat@CPE98fc1165cb6f-CM00159a09b2d0.cpe.net.cable.rogers.com> has joined #aichallenge
2011-12-08T12:52:11  <thestinger> stupid freenode servers :(
2011-12-08T12:55:50  <olexs> gah, some bot on fluxid is using ~3s per turn... game's been running since ages
2011-12-08T12:56:22  <thestinger> yeah
2011-12-08T12:56:32  <thestinger> games take so long now, so I'm unable to test things on tcp
2011-12-08T12:56:48  <pairofdice> How are the other servers?
2011-12-08T12:56:51  <thestinger> Fluxid just needs to get the server to tell bots the turntime is 500ms :P
2011-12-08T12:57:00  <pairofdice> tc-pants and bhickey?
2011-12-08T12:57:38  <BenJackson> bhickey is for lower level competition
2011-12-08T12:58:00  <pairofdice> How low?
2011-12-08T12:58:10  <BenJackson> really low
2011-12-08T12:58:16  <BenJackson> sub-combat
2011-12-08T12:58:26  <pairofdice> Hmph
2011-12-08T12:58:26  <BenJackson> well, people developing new combat can certainly use it
2011-12-08T12:58:34  <BenJackson> but if you're playing there just to provide a bot, provide a weak bot
2011-12-08T13:00:15  <thestinger> hmmmm
2011-12-08T13:00:25  <thestinger> I don't really get how memetix's way works now
2011-12-08T13:00:33  *** Scryer <Scryer!4a645367@gateway/web/freenode/ip.74.100.83.103> has quit IRC (Quit: Page closed)
2011-12-08T13:00:36  <thestinger> I guess he _always_ moves toward the enemy?
2011-12-08T13:01:41  <pairofdice> He masks the tiles on which ants can attack on the next turn
2011-12-08T13:02:03  <pairofdice> just adds up the number of ants that a tile can be attacked by
2011-12-08T13:02:20  <thestinger> yeah, I get that
2011-12-08T13:02:24  <thestinger> I've been doing it all along
2011-12-08T13:02:38  <thestinger> but then he's pretending all his ants are moving towards the enemy, before he decides to do that
2011-12-08T13:02:41  <thestinger> basically
2011-12-08T13:02:44  <pairofdice> oh
2011-12-08T13:02:59  <BenJackson> I don't think so
2011-12-08T13:03:03  <BenJackson> he's pretending all ants move everywhere
2011-12-08T13:03:18  <thestinger> yeah, so lets say he has 2 ants out of range of an enemy
2011-12-08T13:03:21  *** kire <kire!~kire@78-22-125-201.access.telenet.be> has joined #aichallenge
2011-12-08T13:03:26  <thestinger> he's pretending the enemy moves every direction, and his own ants do
2011-12-08T13:03:29  <thestinger> so he can kill the enemy
2011-12-08T13:03:40  <thestinger> but he must be forcing both his ants to then move towards the enemy
2011-12-08T13:03:51  <thestinger> otherwise one might think 'hey!, it's safe!', but it's not
2011-12-08T13:03:58  <thestinger> unless all ants attack if possible
2011-12-08T13:04:06  <pairofdice> Yeah, that's an issue
2011-12-08T13:04:14  <thestinger> I can easily work this into what I already do
2011-12-08T13:04:16  *** Israfel <Israfel!~jeff@99-0-14-82.lightspeed.frokca.sbcglobal.net> has quit IRC (Ping timeout: 240 seconds)
2011-12-08T13:04:25  <thestinger> but I just don't see how it can actually be a complete combat implementation
2011-12-08T13:04:41  <thestinger> or even a big part of it :P
2011-12-08T13:04:45  <pairofdice> It makes it hard to ignore enemy ants
2011-12-08T13:05:06  <thestinger> well, you have to always attack or this way don't work (if I understand it correctly)
2011-12-08T13:05:28  <pairofdice> Pretty much
2011-12-08T13:05:38  *** dvladim <dvladim!~dvladim@85.113.131.65> has quit IRC (Ping timeout: 240 seconds)
2011-12-08T13:05:56  <BenJackson> come on, we're all going to implement it to see what it does...
2011-12-08T13:06:05  <BenJackson> if for no other reason than dozens of people in the forums will
2011-12-08T13:06:10  <BenJackson> so we have to practice against it :)
2011-12-08T13:06:11  <thestinger> I implemented it
2011-12-08T13:06:17  <thestinger> I don't understand how it can possibly work though xD
2011-12-08T13:06:25  <BenJackson> (no one can understand a1k0n's luckily ;-)
2011-12-08T13:06:38  <pairofdice> I've had it for a month on the server now
2011-12-08T13:08:55  *** Apophis <Apophis!~apophis@80-218-68-103.dclient.hispeed.ch> has joined #aichallenge
2011-12-08T13:10:07  <thestinger> oh, now it makes more sense
2011-12-08T13:10:17  <thestinger> or not x
2011-12-08T13:10:19  <thestinger> xD
2011-12-08T13:11:03  <ikaros> :)
2011-12-08T13:11:12  *** olexs1 <olexs1!~Leviathan@drms-590c5f5b.pool.mediaWays.net> has joined #aichallenge
2011-12-08T13:11:33  <thestinger> it's just around 10 lines of code since I already have generic radius utility stuff
2011-12-08T13:11:52  *** Apophis <Apophis!~apophis@80-218-68-103.dclient.hispeed.ch> has quit IRC (Client Quit)
2011-12-08T13:11:58  *** Apophis_ <Apophis_!~apophis@80-218-68-103.dclient.hispeed.ch> has quit IRC (Quit: Verlassend)
2011-12-08T13:12:21  *** Apophis <Apophis!~apophis@80-218-68-103.dclient.hispeed.ch> has joined #aichallenge
2011-12-08T13:13:29  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has quit IRC (Ping timeout: 252 seconds)
2011-12-08T13:14:03  <pairofdice> My bot won't be finished till long after the competition :<
2011-12-08T13:14:07  <ikaros> just do it (c)
2011-12-08T13:14:14  *** olexs <olexs!~Leviathan@drms-4d01513d.pool.mediaWays.net> has quit IRC (Ping timeout: 252 seconds)
2011-12-08T13:14:52  <thestinger> http://sprunge.us/QQgU
2011-12-08T13:14:55  <thestinger> I get something like this now
2011-12-08T13:15:11  <besh> I am back. So did any one understand the method fully ?
2011-12-08T13:15:15  <thestinger> 0 == "SAFE", 1 == "KILL", 2 == "DIE"
2011-12-08T13:15:35  <thestinger> that blob is 2 enemy ants next to each other
2011-12-08T13:15:38  <thestinger> with some of mine around them
2011-12-08T13:15:49  <thestinger> well, mine are on the west
2011-12-08T13:16:43  *** Israfel <Israfel!~jeff@99.0.14.82> has joined #aichallenge
2011-12-08T13:17:38  <thestinger> anyway I don't really get how this turns into combat :)
2011-12-08T13:18:02  <pairofdice> O_o
2011-12-08T13:18:54  <thestinger> like, moving into tiles with 0 (SAFE) could be lethal if your other ants don't move where they are supposed to
2011-12-08T13:19:23  <pairofdice> Yes, but you can control what your ants
2011-12-08T13:19:30  <pairofdice> do
2011-12-08T13:19:39  <thestinger> yeah, but lets say the enemy is on both sides of you
2011-12-08T13:19:44  <thestinger> so if you attack one way, you can beat them
2011-12-08T13:19:49  <thestinger> if you attack the other, you can too
2011-12-08T13:19:54  <thestinger> but if you split up, you will die
2011-12-08T13:19:58  *** anzen <anzen!75fdb999@gateway/web/freenode/ip.117.253.185.153> has joined #aichallenge
2011-12-08T13:20:06  <besh> thestinger: agreed. Moves are coordinated and you can not infer on the outcome of making a move just by looking at influence table.
2011-12-08T13:20:16  <rwest> thestinger: the side with better chances will get most of the weight
2011-12-08T13:20:26  *** bluegaspode <bluegaspode!55b223f9@gateway/web/freenode/ip.85.178.35.249> has quit IRC (Ping timeout: 265 seconds)
2011-12-08T13:20:39  *** anzen <anzen!75fdb999@gateway/web/freenode/ip.117.253.185.153> has quit IRC (Client Quit)
2011-12-08T13:20:49  <thestinger> rwest: ah, so you're pathfinding towards enemies then?
2011-12-08T13:20:58  <thestinger> I don't see how it could work otherwise
2011-12-08T13:21:08  <thestinger> like you need to turn this into an influence map and move towards the enemy
2011-12-08T13:21:26  <thestinger> meh, I don't really get it xD
2011-12-08T13:21:32  <pairofdice> heh
2011-12-08T13:22:28  <rwest> thestinger: you have to diffuse the map
2011-12-08T13:23:10  <rwest> I didn
2011-12-08T13:23:16  <rwest> 't completely read the thread hah
2011-12-08T13:23:29  <rwest> but I assume he is doing a typical influence map
2011-12-08T13:23:36  <rwest> and diffusing the values afterwards
2011-12-08T13:25:04  <besh> i don't think there is diffusion.
2011-12-08T13:25:22  <rwest> ahh, then it is not exactly like mine heh
2011-12-08T13:26:13  <thestinger> rwest: yeah, he just doesn't move into tiles marked as DIE (2 in my map)
2011-12-08T13:26:16  <besh> so did you guys figure out what the best ant is used for ? In other words,how is fighting[p][r][c] determined ?
2011-12-08T13:26:37  <thestinger> sec, I'll explain how I interpreted it
2011-12-08T13:26:38  <thestinger> so
2011-12-08T13:27:04  <thestinger> first of all, I'm putting all this in the Square struct the C++ starter has
2011-12-08T13:27:16  <thestinger> so then I access stuff like grid[row][col].influence[player]
2011-12-08T13:27:37  <thestinger> anyway, for each Square, I make an influence array that holds the number of ants within attackradius+1 for each player
2011-12-08T13:27:53  <thestinger> and a total_influence int with the total number of ants
2011-12-08T13:28:03  <besh> so far so good
2011-12-08T13:28:29  <thestinger> then I make a 'fighting' array (for each player), which is 'total - influence[player]'
2011-12-08T13:28:31  <thestinger> so
2011-12-08T13:28:36  <thestinger> the number of enemies within attackradius+1
2011-12-08T13:29:02  <thestinger> then I go through each tile to decide whether to mark them as DIE or KILL (they are all set to SAFE at the start of the turn)
2011-12-08T13:29:33  <besh> wait
2011-12-08T13:29:48  <besh> He said "enemies = total[row][col] - influence[player][row][col]"
2011-12-08T13:29:52  <thestinger> yeah
2011-12-08T13:30:00  <thestinger> so
2011-12-08T13:30:17  <thestinger> enemies[player] = total - influence[player] for each tile
2011-12-08T13:30:19  <besh> "the best ant being the one that is fighting the least number of other ants"
2011-12-08T13:30:26  <thestinger> yes
2011-12-08T13:30:33  <thestinger> so when I look around each tile (to mark them as DIE or KILL)
2011-12-08T13:30:46  <thestinger> if I find an enemy ant in attackradius+1, I check the # of enemies that ant has
2011-12-08T13:30:55  <besh> so fighting[p][r][c] is not same as number of enemies
2011-12-08T13:31:42  *** zoli1 <zoli1!~zoli@catv-80-98-34-34.catv.broadband.hu> has left #aichallenge ("Konversation terminated!")
2011-12-08T13:31:48  <thestinger> i.e. fighting[0][0][0] would contain the number of enemy ants an ant at (0,0) belonging to player 0 would be fighting.
2011-12-08T13:31:51  <thestinger> that's what he says
2011-12-08T13:31:53  <thestinger> so I just did that
2011-12-08T13:32:04  <thestinger> I find the best ant based on that
2011-12-08T13:32:14  <thestinger> 'best ant' meaning enemy with the lowest weakness
2011-12-08T13:32:21  <thestinger> just like running a combat evaluation
2011-12-08T13:32:56  <besh> yes that second statement makes sense.
2011-12-08T13:33:24  <thestinger> I don't really think I interpreted what he said in the right way
2011-12-08T13:33:37  <thestinger> anyway
2011-12-08T13:33:42  *** dr-_ <dr-_!~dr-@ble59-1-82-66-133-150.fbx.proxad.net> has joined #aichallenge
2011-12-08T13:33:46  <besh> I thought we are going to select our ant with the least number of enemy attacks.
2011-12-08T13:33:59  <thestinger> fighting[row][col][player] is the number of enemies (of that player) within attackradius+1
2011-12-08T13:34:09  <thestinger> besh: yeah, I do that with the fighting data later
2011-12-08T13:34:19  <thestinger> so after I calculate all that stuff
2011-12-08T13:34:22  <thestinger> I go through each tile
2011-12-08T13:34:42  <thestinger> I get the fighting value of that tile (for player 0)
2011-12-08T13:34:52  <thestinger> then I look at the tiles around it (within attackradius+1)
2011-12-08T13:35:16  <thestinger> if there's an enemy ant, I check the value for it (# of enemies it has)
2011-12-08T13:35:43  *** dr- <dr-!~dr-@ble59-1-82-66-133-150.fbx.proxad.net> has quit IRC (Ping timeout: 268 seconds)
2011-12-08T13:35:52  <thestinger> then I compare to the value I got for the tile I'm checking - it's equal, I mark it KILL - if it's greater, I mark it DIE and break out of the loop
2011-12-08T13:36:55  <besh> Ok thanks it is getting clearer now.
2011-12-08T13:37:03  <thestinger> well, that's how I interpreted it
2011-12-08T13:37:14  <thestinger> I think my interpretation must be wrong because this doesn't seem very useful :P
2011-12-08T13:37:42  <besh> haha... it must be because his bot is very strong ;)
2011-12-08T13:38:47  <besh> My combat scheme which I think a lot other use is simple
2011-12-08T13:39:01  <besh> I just park my ants on attack_radius + 2 of enemy ants
2011-12-08T13:39:13  <besh> then attack when I outnumber the enemy ant
2011-12-08T13:39:31  <thestinger> so does it work if there's 1 enemy
2011-12-08T13:39:38  <thestinger> and you have 1 ant on opposite sides?
2011-12-08T13:40:19  <thestinger> meh, maybe I'll try the debug visualizer and take a screen of this crap :P
2011-12-08T13:43:43  <BenJackson> wow I played a burst of games
2011-12-08T13:43:56  <BenJackson> 5 between 3:16am and 3:23am
2011-12-08T13:44:09  <BenJackson> previous gap was about 24h
2011-12-08T13:44:19  <BenJackson> net skill change was -0.05
2011-12-08T13:44:41  <besh> yes it works well but sometimes can suicide.
2011-12-08T13:44:55  <besh> A little correction there. I park them at attack_radius + 1
2011-12-08T13:45:04  <besh> and also attack_radius + 2
2011-12-08T13:45:36  <besh> In 2 vs 1 it can never loose but doesn't mean it can always win.
2011-12-08T13:46:22  <BenJackson> since my combat just maximizes certain heuristics I often don't understand what it does
2011-12-08T13:46:27  <thestinger> oh
2011-12-08T13:46:29  <thestinger> I think this works :D
2011-12-08T13:46:31  <BenJackson> or not without significant thought
2011-12-08T13:46:52  <thestinger> http://ompldr.org/vYm5jZw
2011-12-08T13:46:58  <thestinger> so that orange ant near the enemies
2011-12-08T13:47:05  <thestinger> if it moves down (into blue) it can trade 1 for 1
2011-12-08T13:47:06  <BenJackson> it will sort of suck if after seeing everyone implement the same strategy function and clump up in the ranks now everyone implements the same combat
2011-12-08T13:47:14  <thestinger> if it moves out of the red/blue ranges, it will be safe
2011-12-08T13:47:40  <pairofdice> Why is the blue only south
2011-12-08T13:47:53  <BenJackson> blue is trade
2011-12-08T13:47:56  <BenJackson> if it moves east it can escape
2011-12-08T13:48:42  <thestinger> yeah, this is somewhat different than what I do now
2011-12-08T13:48:46  <thestinger> I think I could easily mix them together
2011-12-08T13:48:54  <thestinger> but I don't think just doing it this way will work
2011-12-08T13:49:15  <Fluxid> oh boy, i'm reading backscroll
2011-12-08T13:49:23  <thestinger> Fluxid: just look at the pic :P
2011-12-08T13:49:24  <Fluxid> and i do stuff totally different
2011-12-08T13:49:29  <thestinger> yeah
2011-12-08T13:49:35  <Fluxid> but at the same way totally inefficient
2011-12-08T13:49:45  <Fluxid> time*
2011-12-08T13:49:50  <thestinger> atm I come up with possible enemy locations
2011-12-08T13:49:59  <thestinger> a simple version is just unmoving enemies, and enemies moving towards me
2011-12-08T13:50:18  <thestinger> then I generate moves that optimize a "score" based on those enemy locations (I assume they will do the best thing, so I raise the minimum)
2011-12-08T13:50:27  <thestinger> so like a payoff matrix, but only for enemy moves
2011-12-08T13:51:43  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has joined #aichallenge
2011-12-08T13:51:51  <BenJackson> actually this combat is pretty much the exact inverse of what I do
2011-12-08T13:52:01  <BenJackson> this method looks at how enemies could possibly threaten you
2011-12-08T13:52:12  <BenJackson> and I look at how I could possibly threaten enemies
2011-12-08T13:52:50  <Fluxid> those ideas look neat bu i think i'll stick with what i do
2011-12-08T13:53:06  <antimatroid> thestinger: if you're only considering 1 set of moves from your opponents it's more like a decision matrix :P
2011-12-08T13:53:29  <thestinger> antimatroid: no, I consider n sets of enemy moves
2011-12-08T13:53:38  <thestinger> but I only test 1 set of moves against them and just keep optimizing it
2011-12-08T13:53:54  <thestinger> so it's like a payoff matrix but I don't actually keep it around
2011-12-08T13:54:14  <thestinger> anyway it doesn't work very well
2011-12-08T13:55:23  <thestinger> http://aichallenge.org/visualizer.php?game=171150&user=2255
2011-12-08T13:55:37  <thestinger> you can see that it gets the basics okay, but it falls apart for big battles
2011-12-08T13:55:47  <BenJackson> which one is he?
2011-12-08T13:55:55  <thestinger> because I don't form lines
2011-12-08T13:55:55  <BenJackson> oh you mean YOU
2011-12-08T13:56:06  <thestinger> and I don't really try to flank the enemy, even when I outnumber them
2011-12-08T13:56:29  <thestinger> the lines that do form are just because my ants try to get around each other, but that's limited to ~3 tiles away
2011-12-08T13:56:32  <BenJackson> neither do I, although it happens as an emergent behavior
2011-12-08T13:57:00  <BenJackson> I used to use "step around" avoidance but it resulted in lines
2011-12-08T13:57:10  <BenJackson> it's neat to see the current code moves 3-abreast snakes through mazes and stuff
2011-12-08T13:57:40  <BenJackson> the other problem with 'stepping around' is that the ants that would step out of line would sometimes step right back in to be closer, followed by stepping out, etc
2011-12-08T13:57:51  <Minthos> I saw one bot whose ants were just a flowing mass, like liquid
2011-12-08T13:58:25  <antimatroid> thestinger: have you considered path finding to enemies not over friends to try and get them to surround the enemy?
2011-12-08T13:59:33  <antimatroid> or vice versa
2011-12-08T13:59:40  <antimatroid> from enemies to friends but not over friends
2011-12-08T13:59:57  <thestinger> yeah
2011-12-08T14:00:06  <thestinger> I have a few attempts in branches atm
2011-12-08T14:00:19  <thestinger> such as pretending a layer thicker than 2 of my ants is not passable
2011-12-08T14:00:21  <antimatroid> i haven't touched my bot in a few days
2011-12-08T14:00:44  <antimatroid> been busy with other stuff, hopefully i have time again on the weekend, although i need to fix up the maze map gen before my bot again too
2011-12-08T14:00:54  <thestinger> my bot is basically fine until there's a big battle, where it will just form a ball instead of a line
2011-12-08T14:01:20  <antimatroid> because they're all either agressive or defensive?
2011-12-08T14:01:45  *** UncleVasya <UncleVasya!kvirc@128-124-31-21.dialup.umc.net.ua> has joined #aichallenge
2011-12-08T14:01:49  <thestinger> well, because they don't actually chase the enemy
2011-12-08T14:02:00  *** bluegaspode <bluegaspode!55b223f9@gateway/web/freenode/ip.85.178.35.249> has joined #aichallenge
2011-12-08T14:02:04  <antimatroid> i think my move selection is quite different to yours
2011-12-08T14:02:12  <antimatroid> but i'm not willing to share how i'm picking out moves :P
2011-12-08T14:02:22  <thestinger> basically, I have efficient movement to a target
2011-12-08T14:02:42  <thestinger> so, they'll usually form a 3 or 4 width march towards it
2011-12-08T14:02:49  *** Palmik <Palmik!~palmik@man-208-200.pks.muni.cz> has quit IRC (Remote host closed the connection)
2011-12-08T14:03:10  <thestinger> but then they don't really spread out further when there's a battle
2011-12-08T14:03:18  *** Palmik <Palmik!~palmik@man-208-200.pks.muni.cz> has joined #aichallenge
2011-12-08T14:03:21  <thestinger> just as much as needed to get efficient movement (assuming there was no enemy)
2011-12-08T14:04:24  <thestinger> meh, I think I misunderstood this new way of doing combat xD
2011-12-08T14:04:43  <thestinger> http://ompldr.org/vYm5kOQ
2011-12-08T14:04:47  *** bhasker <bhasker!~bhasker@99-162-151-97.lightspeed.sntcca.sbcglobal.net> has quit IRC (Ping timeout: 240 seconds)
2011-12-08T14:04:49  <thestinger> red == "DIE"
2011-12-08T14:04:56  <thestinger> so... it doesn't work like this
2011-12-08T14:06:20  <thestinger> because the "best" enemy ant doesn't have any enemies within attackradius+1
2011-12-08T14:07:27  <antimatroid> lol, you're still using attackradius +1?
2011-12-08T14:07:30  <antimatroid> partition them properly!
2011-12-08T14:07:39  <thestinger> no
2011-12-08T14:07:45  <thestinger> this is to implement this http://forums.aichallenge.org/viewtopic.php?f=24&t=2083
2011-12-08T14:07:50  <thestinger> I'm trying to understand it
2011-12-08T14:08:21  <thestinger> I could just get attackradius with the ant moved in each possible direction
2011-12-08T14:08:33  <thestinger> but there's no point doing it properly if it's not going to work at all :P
2011-12-08T14:09:11  <antimatroid> hehe, that's a lot like what i do, except they missed a trick i'm using to get slightly more accurate info
2011-12-08T14:09:42  <antimatroid> i use that kind of information to pick out offensive and defensive moves for my opponents and i then put it through a game tree
2011-12-08T14:11:25  <UncleVasya> Offencive is where sum of  summes of distances every enemy to every your ant is minimal and defencive vise versa?
2011-12-08T14:16:01  <antimatroid> thestinger: http://pastebin.com/a0tstdX5
2011-12-08T14:16:08  <antimatroid> that's a hint towards improving the accuracy of your info
2011-12-08T14:16:18  <antimatroid> only 1 enemy can really battle a if he remains statis
2011-12-08T14:16:20  <antimatroid> static*
2011-12-08T14:17:15  <thestinger> well, I know a better way would be tiles within attackradius of the tile the enemy ant is in, and also the tiles they could move to
2011-12-08T14:18:01  <antimatroid> yeah, i just do min of those two numbers :P
2011-12-08T14:18:10  <antimatroid> unique locations battled from and unique enemies
2011-12-08T14:19:44  <antimatroid> you only need to work out this information for locations that can be moved into as well obviously
2011-12-08T14:21:05  <antimatroid> and my state stores for each location the locations that (i) could battle it if it doesn't move and it's enemies do(don't) (ii) could battle if it does(n't) and it's enemies do(don't) move
2011-12-08T14:21:50  <antimatroid> you can set them up with translation values in turn 0 then just update as you receive water info :P
2011-12-08T14:26:06  *** vn971 <vn971!~vasya@77.232.28.15> has joined #aichallenge
2011-12-08T14:26:16  <vn971> test
2011-12-08T14:26:22  <rwest> ies
2011-12-08T14:30:13  <thestinger> antimatroid: http://ompldr.org/vYm5kdw
2011-12-08T14:30:28  <thestinger> blue == possible trade, red == possible death
2011-12-08T14:31:24  <antimatroid> your info isn't right :P
2011-12-08T14:31:32  <antimatroid> oh wait
2011-12-08T14:31:33  <antimatroid> ignore me
2011-12-08T14:31:48  <antimatroid> well, the top seems wrong
2011-12-08T14:31:50  <antimatroid> but not important
2011-12-08T14:31:58  <antimatroid> the top left red square?
2011-12-08T14:32:04  <antimatroid> how is that a possible death? :P
2011-12-08T14:32:17  <thestinger> if the ant moves north
2011-12-08T14:32:22  <thestinger> they can attack 2 forwards and 1 west
2011-12-08T14:32:22  <UncleVasya> antimatroid: don't you know where mcstar's gone?
2011-12-08T14:32:27  <antimatroid> oh, by the time it gets there?
2011-12-08T14:32:37  <antimatroid> UncleVasya: nope? how long has he been gone?
2011-12-08T14:32:39  <antimatroid> @last seen mcstar
2011-12-08T14:32:39  <contestbot> antimatroid: (last [--{from,in,on,with,without,regexp} <value>] [--nolimit]) -- Returns the last message matching the given criteria. --from requires a nick from whom the message came; --in requires a channel the message was sent to; --on requires a network the message was sent on; --with requires some string that had to be in the message; --regexp requires a regular expression the message must (1 more message)
2011-12-08T14:32:53  <UncleVasya> @seen mcstar
2011-12-08T14:32:53  <contestbot> UncleVasya: mcstar was last seen in #aichallenge 1 week, 3 days, 20 hours, 26 minutes, and 56 seconds ago: <mcstar> i dont agree
2011-12-08T14:33:00  <antimatroid> hmmmm :\
2011-12-08T14:33:05  <thestinger> antimatroid: well, it's "possible death once the enemies move"
2011-12-08T14:33:19  <thestinger> or possible trade once they move
2011-12-08T14:33:34  <antimatroid> thestinger: so the info is for when your ant could get to the location?
2011-12-08T14:33:49  <thestinger> well, I really only need to calculate this for tiles next to my ants
2011-12-08T14:33:53  <antimatroid> do you get my example from before right?
2011-12-08T14:34:07  <antimatroid> http://pastebin.com/a0tstdX5
2011-12-08T14:34:12  <thestinger> yeah, one of the blue ants couldn't move down
2011-12-08T14:34:15  <antimatroid> ie. if you move 'a' down is it blue above him?
2011-12-08T14:34:18  <thestinger> and they can't both move closer to a
2011-12-08T14:34:19  <antimatroid> yeah
2011-12-08T14:34:21  <thestinger> without colliding
2011-12-08T14:34:29  <antimatroid> it took me a while to work out how to get that info
2011-12-08T14:35:56  *** kire <kire!~kire@78-22-125-201.access.telenet.be> has quit IRC (Remote host closed the connection)
2011-12-08T14:36:16  *** JorgeB <JorgeB!~textual@209.66.74.34> has quit IRC (Ping timeout: 240 seconds)
2011-12-08T14:39:33  *** hkraal_ <hkraal_!4dfa050f@gateway/web/freenode/ip.77.250.5.15> has joined #aichallenge
2011-12-08T14:39:58  *** UncleVasya <UncleVasya!kvirc@128-124-31-21.dialup.umc.net.ua> has quit IRC ()
2011-12-08T14:52:45  *** AntDroid <AntDroid!70c6400f@gateway/web/freenode/ip.112.198.64.15> has quit IRC (Ping timeout: 265 seconds)
2011-12-08T14:56:01  <besh> antimatroid: I take care of your example problem in my bot.
2011-12-08T14:56:08  *** McLeopold <McLeopold!~Scott@c-98-247-248-39.hsd1.wa.comcast.net> has joined #aichallenge
2011-12-08T14:56:09  <antimatroid> :)
2011-12-08T14:56:22  <McLeopold> antimatroid: http://paste.aichallenge.org/NidJk/?row ... 14&turn=33
2011-12-08T14:56:29  <besh> ants on attack_radius + 1 are allowed to move only if they outnumber the opponent
2011-12-08T14:56:58  <thestinger> if you're doing it that way you need attackradius+2
2011-12-08T14:57:11  <thestinger> because the enemy can move towards you, in addition to you moving towards them
2011-12-08T14:57:35  <besh> attack_radius + 2 ants are always moved closer unless it is always suicide ofcourse
2011-12-08T14:57:45  <thestinger> ah
2011-12-08T14:58:03  <McLeopold> http://paste.aichallenge.org/NidJk/?row=3&col=14&turn=33
2011-12-08T14:58:07  <McLeopold> that's better
2011-12-08T14:58:21  <antimatroid> what am i looking at though?
2011-12-08T14:58:24  <McLeopold> I'm trying to implement Memetix's combat
2011-12-08T14:58:41  <McLeopold> the highlighted ant is mine
2011-12-08T14:58:52  <McLeopold> it thinks it can safely go south, but it can't
2011-12-08T14:59:00  <Fluxid> http://paste.aichallenge.org/GpWxL/ go on, laugh at me
2011-12-08T14:59:08  <antimatroid> it could when it did
2011-12-08T14:59:37  <antimatroid> you mean west?
2011-12-08T14:59:43  <McLeopold> no, south
2011-12-08T14:59:54  <antimatroid> :s
2011-12-08T15:00:15  <McLeopold> I think the ant at 6,12 is throwing it off
2011-12-08T15:00:18  <besh> fluixd: what went wrong there? your combat is too cautious
2011-12-08T15:00:29  <thestinger> McLeopold: draw the status for all the tiles
2011-12-08T15:00:31  *** vn971 <vn971!~vasya@77.232.28.15> has left #aichallenge
2011-12-08T15:00:34  <thestinger> not just the ones next to your ants
2011-12-08T15:00:39  <thestinger> makes it easier to see what you're doing :P
2011-12-08T15:00:39  <antimatroid> it realises that the enemy could be smarter
2011-12-08T15:00:40  <McLeopold> ok
2011-12-08T15:00:50  <pairofdice> McLeopold: Yeah, it thinks it has backup when it doesn't
2011-12-08T15:00:56  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has quit IRC (Ping timeout: 252 seconds)
2011-12-08T15:00:57  <McLeopold> I know
2011-12-08T15:00:59  <pairofdice> possibly
2011-12-08T15:01:06  <McLeopold> but, I don't think I'm updating stuff right
2011-12-08T15:01:38  * antimatroid doesn't see an ant at 6 12?
2011-12-08T15:01:38  <Fluxid> besh: everything. I have in logs stuff like "I may kill something when going 2" but no ant actually go in direction which kills anything
2011-12-08T15:02:04  <antimatroid> oh wait, wrong turn :P
2011-12-08T15:02:12  <McLeopold> 2 of my ants *could* influence 4,14, and only 1 enemy *could* influence 4,14, but it should still be KILL
2011-12-08T15:02:15  <thestinger> Fluxid: I'm just throwing out all my combat code for the 5th time and starting over :P
2011-12-08T15:02:51  <pairofdice> I wonder if I get an update before the end
2011-12-08T15:03:00  <Fluxid> thestinger: pasted the map here: http://paste.aichallenge.org/DVwbi/
2011-12-08T15:03:13  <antimatroid> McLeopold: what I do is use that information to pick out offensive and defensive moves for each ant in my battle partition, those for my ants constitute 2 moves in a game tree and those of my enemies are 2 moves for them in a game tree, then i run minimax on that to work out whether to be offensive or defensive
2011-12-08T15:03:15  <pairofdice> Would like to nudge my bot from #150 to top 100
2011-12-08T15:04:07  <Fluxid> someone is posting lots of pastes
2011-12-08T15:04:13  <antimatroid> i would suggest that one could perhaps try some hill climbing like heuristics once you've picked your move to improve it
2011-12-08T15:04:16  <amstan_> McLeopold: how did you get rid of the spam on antpaste?
2011-12-08T15:04:29  <antimatroid> has anyone tried that last suggestion?
2011-12-08T15:04:33  <McLeopold> amstan_: I just looked for urls and denied the paste
2011-12-08T15:04:47  <amstan_> doesn't the json have urls in it?
2011-12-08T15:04:55  <McLeopold> yes
2011-12-08T15:04:58  <McLeopold> I allow those
2011-12-08T15:05:04  <amstan_> cool
2011-12-08T15:05:15  <McLeopold> not really, spammers are just stupid, or lazy
2011-12-08T15:05:56  <McLeopold> the php paster is really anoying...
2011-12-08T15:06:15  <thestinger> Fluxid: my combat attacks in that map
2011-12-08T15:06:46  <thestinger> wait nvm I have combat disabled because I'm testing this thing - it still attacks with it enabled though
2011-12-08T15:07:00  <Fluxid> :|
2011-12-08T15:07:20  <thestinger> the 4 ants at the front move forwards
2011-12-08T15:07:26  <Fluxid> what's even more funny, my bot takes all the time every second turn on this time
2011-12-08T15:07:33  <thestinger> the 2 ants near the 1 enemy in the easy move forwards and kill it
2011-12-08T15:07:38  <Fluxid> like, one turn takes 10ms, and second 467ms
2011-12-08T15:07:40  <Fluxid> wtf
2011-12-08T15:07:50  *** delt0r_ <delt0r_!~bob@178-190-41-115.adsl.highway.telekom.at> has quit IRC (Ping timeout: 248 seconds)
2011-12-08T15:07:50  <thestinger> and with the bottom part, one ant waits until the other gets there and they kill it
2011-12-08T15:07:54  <thestinger> my bot handles unmoving enemies fine
2011-12-08T15:08:00  <thestinger> it's the moving ones that screw it up
2011-12-08T15:08:08  <thestinger> when they move sideways instead of towards me
2011-12-08T15:09:54  *** Anilm3 <Anilm3!~anil@4.Red-79-146-241.dynamicIP.rima-tde.net> has quit IRC (Quit: Lost terminal)
2011-12-08T15:12:53  <thestinger> http://ompldr.org/vYm5lcg well, this does something
2011-12-08T15:12:56  <thestinger> I'm not sure what though
2011-12-08T15:13:13  <pairofdice> looks funky
2011-12-08T15:13:20  *** Redgis <Redgis!~Nex@2a01:e34:ee3b:aa40:ccb:8a1d:94b0:4c6> has joined #aichallenge
2011-12-08T15:14:43  *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge
2011-12-08T15:16:40  *** Kurnevsky <Kurnevsky!d91529cd@gateway/web/freenode/ip.217.21.41.205> has joined #aichallenge
2011-12-08T15:20:43  *** delt0r_ <delt0r_!~bob@91-113-94-210.adsl.highway.telekom.at> has joined #aichallenge
2011-12-08T15:21:50  *** vn971 <vn971!~vasya@77.232.28.15> has joined #aichallenge
2011-12-08T15:22:22  <Kurnevsky> Tell me please, where should I place my files so my bot can read them?
2011-12-08T15:22:54  <McLeopold> at the root of the zip
2011-12-08T15:23:15  <McLeopold> your bot's current directory will always be that
2011-12-08T15:23:47  <Kurnevsky> ok, thank you
2011-12-08T15:24:47  <thestinger> oh well, this method doesn't really work for me but it compliments my existing combat code well
2011-12-08T15:24:57  <thestinger> I can generate a better initial set of offensive moves
2011-12-08T15:25:16  <thestinger> http://ompldr.org/vYm5mOQ can just completely ignore moves into red zones
2011-12-08T15:25:20  *** Accoun <Accoun!~1@static-213-88-39-97.chebnet.ru> has quit IRC (Remote host closed the connection)
2011-12-08T15:25:24  <thestinger> and do my combat the old way, but with a better start
2011-12-08T15:25:33  *** Accoun <Accoun!~1@173.195.0.224> has joined #aichallenge
2011-12-08T15:26:06  <Wraithan> I am looking forward to combat being my problem
2011-12-08T15:27:50  <McLeopold> okay, I was updating the status wrong http://paste.aichallenge.org/puzdB/
2011-12-08T15:28:26  <thestinger> McLeopold: so you just don't move into DIE zones?
2011-12-08T15:28:28  <McLeopold> it makes me smile to watch the replay with visualizations and see the ants lock on to the hill :)
2011-12-08T15:28:44  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has joined #aichallenge
2011-12-08T15:29:22  <McLeopold> thestinger: only when exploring and attacking
2011-12-08T15:29:29  <McLeopold> when gathering food, I go for it
2011-12-08T15:33:31  *** pairofdice <pairofdice!~pod@dsl-hkibrasgw2-ff67c300-50.dhcp.inet.fi> has quit IRC (Quit: leaving)
2011-12-08T15:33:37  <McLeopold> okay, who's on what tcp server?
2011-12-08T15:33:51  *** Accoun <Accoun!~1@173.195.0.224> has quit IRC ()
2011-12-08T15:33:57  <thestinger> I think most people are on Fluxid's tcp server
2011-12-08T15:34:09  <thestinger> http://ants.fluxid.pl/ranking
2011-12-08T15:34:18  <McLeopold> that's the good one?
2011-12-08T15:34:30  <McLeopold> anyone playing on the crappy one? :)
2011-12-08T15:34:32  <thestinger> well, they're all good - but no one plays on the others :P
2011-12-08T15:34:38  *** Hexren <Hexren!~hexren@46.59.196.52> has joined #aichallenge
2011-12-08T15:34:46  <McLeopold> I'll give it a shot
2011-12-08T15:35:16  *** Hexren <Hexren!~hexren@46.59.196.52> has left #aichallenge
2011-12-08T15:35:44  *** vn971 <vn971!~vasya@77.232.28.15> has left #aichallenge
2011-12-08T15:35:59  <besh> someone do something about the slow bots on tcp :(
2011-12-08T15:36:40  <thestinger> Fluxid: change the turntime reported to the bots to 500ms :)
2011-12-08T15:38:48  <besh> Yes. Then the server could wait for a response for 5000ms if it wants. But the bot the should know it only has 500ms
2011-12-08T15:39:22  <McLeopold> does the tcpclient report status and turns?
2011-12-08T15:39:45  *** Accoun <Accoun!~1@static-213-88-39-97.chebnet.ru> has joined #aichallenge
2011-12-08T15:39:47  <McLeopold> yep
2011-12-08T15:40:01  <McLeopold> yeah! timeout!
2011-12-08T15:45:31  <ronchaine> ...if someone has a faintest clue what is going on here, please help.
2011-12-08T15:45:34  <ronchaine> http://pastebin.com/d7SyzCtf
2011-12-08T15:45:54  <ronchaine> assert OK, call function, same assert -> fail.
2011-12-08T15:49:42  *** heinrich5991 <heinrich5991!~hein5991@unaffiliated/heinrich5991> has quit IRC (Excess Flood)
2011-12-08T15:50:30  <besh> ronchaine: are u sure about what you saw? It should behave the same as long as the same value is passed
2011-12-08T15:50:38  <ronchaine> besh: Positive
2011-12-08T15:50:43  <ronchaine> That is why I'm so baffled.
2011-12-08T15:50:59  <g0llum> ronchaine, yo're shure the code yyou pasted is the only call to getloc ?
2011-12-08T15:52:15  *** heinrich5991 <heinrich5991!~hein5991@unaffiliated/heinrich5991> has joined #aichallenge
2011-12-08T15:52:51  <ronchaine> No, there are others, but if I comment out that one call, it won't crash.
2011-12-08T15:53:53  <besh> what value does dir have when it fails ?
2011-12-08T15:55:08  <ronchaine> a sec, I'll check.
2011-12-08T15:56:21  <besh> if it is a garbage value then it is in the "else" block.
2011-12-08T15:56:23  <g0llum> case of 'can't find my way home', lol
2011-12-08T15:57:16  *** Kurnevsky <Kurnevsky!d91529cd@gateway/web/freenode/ip.217.21.41.205> has left #aichallenge
2011-12-08T15:58:12  *** bluegaspode <bluegaspode!55b223f9@gateway/web/freenode/ip.85.178.35.249> has quit IRC (Quit: Page closed)
2011-12-08T15:58:57  *** Palmik <Palmik!~palmik@man-208-200.pks.muni.cz> has quit IRC (Remote host closed the connection)
2011-12-08T16:00:36  *** epicmonkey <epicmonkey!~epicmonke@178.130.5.103> has joined #aichallenge
2011-12-08T16:01:05  <BenJackson> McLeopold: http://paste.aichallenge.org/NidJk/?row=3&col=14&turn=33
2011-12-08T16:01:10  <BenJackson> that doesn't look like ANY combat
2011-12-08T16:01:16  <BenJackson> every meeting was a 1-1 trade
2011-12-08T16:03:06  <ronchaine> besh: It's "1"
2011-12-08T16:05:27  <besh> well that assert can not fail with a value of 1.
2011-12-08T16:05:46  <ronchaine> ...yeah, when it comes to the function, it is suddenly 87.
2011-12-08T16:06:18  <BenJackson> Fluxid: http://paste.aichallenge.org/xH1yL/
2011-12-08T16:06:29  <ronchaine> Constantly 87, even.
2011-12-08T16:07:54  *** heinrich5991 <heinrich5991!~hein5991@unaffiliated/heinrich5991> has quit IRC (Excess Flood)
2011-12-08T16:08:21  <besh> try intitializing d to 0 i.e int d = 0;
2011-12-08T16:08:48  <besh> line 3;
2011-12-08T16:09:44  *** heinrich5991 <heinrich5991!~hein5991@unaffiliated/heinrich5991> has joined #aichallenge
2011-12-08T16:10:55  *** ltriant <ltriant!~ltriant@lithium.mailguard.com.au> has joined #aichallenge
2011-12-08T16:11:26  <ronchaine> Tried, didn't help.  Also, if it doesn't hit any of the NSEW's (which initialise the value) it should fail in the assert(0) -block.
2011-12-08T16:12:50  *** hkraal_ <hkraal_!4dfa050f@gateway/web/freenode/ip.77.250.5.15> has quit IRC (Quit: Page closed)
2011-12-08T16:13:39  <BenJackson> ronchaine: I would guess that getLoc is being inlined (good) and the two identical asserts are being folded into one
2011-12-08T16:13:43  <besh> maybe you should check for other getloc() calls as gollum suggested
2011-12-08T16:13:47  <BenJackson> and the debug information happens to point at the getLoc one
2011-12-08T16:13:56  <BenJackson> try with -O0
2011-12-08T16:13:59  *** tobym <tobym!~tobym@pool-71-183-248-125.nycmny.fios.verizon.net> has quit IRC (Remote host closed the connection)
2011-12-08T16:14:05  <BenJackson> then you will probably see it assert on the expected line
2011-12-08T16:14:44  <besh> it is in debug mode right ?
2011-12-08T16:14:48  <ronchaine> Yeah.
2011-12-08T16:14:50  <besh> no optimization
2011-12-08T16:14:56  <ronchaine> O.o
2011-12-08T16:15:05  <ronchaine> Just tested, with no optimisation it works.
2011-12-08T16:15:13  <thestinger> do you have warnings turned up?
2011-12-08T16:15:15  *** tobym <tobym!~tobym@pool-71-183-248-125.nycmny.fios.verizon.net> has joined #aichallenge
2011-12-08T16:15:24  <BenJackson> it works works or it asserts on line "16"?
2011-12-08T16:15:34  <thestinger> bugs caused by optimization are pretty much always caught by a warning
2011-12-08T16:16:03  <BenJackson> they're pretty much always mem smashing/initialization bugs too
2011-12-08T16:16:23  *** nplus <nplus!~chatzilla@unaffiliated/simont> has quit IRC (Quit: ChatZilla 0.9.87-rdmsoft [XULRunner 1.9.0.17/2009122204])
2011-12-08T16:16:50  <Minthos> use valgrind
2011-12-08T16:17:02  <thestinger> libmudflap is better than valgrind :)
2011-12-08T16:17:22  <Minthos> better than valgrind? blasphemy! :p
2011-12-08T16:17:23  <ikaros> any hint on how much time i should reserver for sending outputs? .. this seems to vary pretty much
2011-12-08T16:17:30  <thestinger> Minthos: much better, it's built into your program
2011-12-08T16:17:31  <ikaros> reserve*
2011-12-08T16:17:33  <thestinger> bounds checks for every array
2011-12-08T16:17:34  <BenJackson> it's 10ms per 100 lines
2011-12-08T16:17:41  <BenJackson> the engine enforces that
2011-12-08T16:17:49  <thestinger> int a[4]; a[5] = 5;
2011-12-08T16:17:53  <BenJackson> but you can send output early and be processing during the 10ms too
2011-12-08T16:17:53  <ikaros> hmm
2011-12-08T16:17:54  <thestinger> valgrind doesn't catch that, libmudflap does
2011-12-08T16:18:14  <thestinger> well valgrind might if it actually accesses off the end of the stack
2011-12-08T16:18:15  <ikaros> i buffered all outputs in a stringstream and send it completely at the end
2011-12-08T16:18:37  <BenJackson> not a good idea
2011-12-08T16:18:38  <thestinger> ikaros: the engine pauses for 10ms between every... 50 lines I think
2011-12-08T16:18:39  <thestinger> maybe 100 lines
2011-12-08T16:18:43  <thestinger> don't buffer it
2011-12-08T16:18:48  <BenJackson> if you know what you want to do, do it early if you have >100 bots
2011-12-08T16:18:56  <ikaros> hm ok i did that because i already had timeouts befor
2011-12-08T16:18:59  <ikaros> which was strange
2011-12-08T16:19:24  <ikaros> set turntime to 50ms and let 10% for output
2011-12-08T16:19:28  <thestinger> so my new combat code is different
2011-12-08T16:19:30  <ikaros> but it seemed to low
2011-12-08T16:19:30  <thestinger> not sure if it's better
2011-12-08T16:19:43  <thestinger> maybe I just have 1000 bugs to fix
2011-12-08T16:20:12  <Minthos> same here.. I fixed last night's bug today
2011-12-08T16:20:21  <Minthos> and made one ore more new ones
2011-12-08T16:20:32  <Minthos> so now my bot is mental again
2011-12-08T16:20:51  <Minthos> s/ore/or/
2011-12-08T16:21:02  <besh> I do the same thning as Ikaros. What is wrong with buffering ?
2011-12-08T16:21:14  <McLeopold> BenJackson: I had updated the status wrong.  And it's not really combat yet, just combat avoidance.
2011-12-08T16:22:00  <ikaros> keep getting timeouts.. even with 10ms reserved for output.. very strange
2011-12-08T16:22:07  <ikaros> and there are only 15 ants or so
2011-12-08T16:22:12  <ikaros> meaning 15 lines..
2011-12-08T16:22:19  <BenJackson> how is 15 ants taking 500ms anyway?
2011-12-08T16:22:21  <besh> thestinger: is it better to send each move as you get it ?
2011-12-08T16:22:23  <Minthos> buffering throws away computation time because the engine is slow at reading bot commands
2011-12-08T16:22:35  <thestinger> besh: I dunno, it doesn't matter for me
2011-12-08T16:22:39  <ikaros> BenJackson, i lowered turntime to 100ms
2011-12-08T16:22:48  <ikaros> and i test how many calculations i can do
2011-12-08T16:22:48  <thestinger> I output all moves near the end because I keep them around to change them with combat code
2011-12-08T16:22:56  <ikaros> thats why it uses 90% time for simulation
2011-12-08T16:23:00  <thestinger> but it would be better just to output them right away I guess
2011-12-08T16:23:19  <McLeopold> just before I timed out, I had the highest ant count \o/ http://ants.fluxid.pl/replay.19470
2011-12-08T16:24:03  <BenJackson> wow I wonder if the "_15lookahead" guy is really etc
2011-12-08T16:24:28  <BenJackson> guess I've got to implement this new combat thing to see it in action
2011-12-08T16:24:29  <thestinger> no way
2011-12-08T16:24:43  <thestinger> it'
2011-12-08T16:24:43  <BenJackson> thestinger: oh you'd know allll about fake suffixes
2011-12-08T16:24:52  <thestinger> it's probably version 15, with lookahead
2011-12-08T16:24:54  <BenJackson> strcat_misdirection
2011-12-08T16:25:07  <thestinger> lol :)
2011-12-08T16:25:19  <BenJackson> watch out for bj_complete_endgame_database
2011-12-08T16:25:23  <besh> Minthos: we are talking about writing your moves. Is it better to write each move one by one as you get it doing computation in between OR write them all at the end ?
2011-12-08T16:25:32  <Minthos> besh: the first
2011-12-08T16:25:47  *** bluegaspode <bluegaspode!55b223f9@gateway/web/freenode/ip.85.178.35.249> has joined #aichallenge
2011-12-08T16:25:49  <BenJackson> I do mine at the end for historical reasons
2011-12-08T16:25:55  <BenJackson> but then I've never timed out
2011-12-08T16:25:59  <BenJackson> and I don't even watch the time
2011-12-08T16:26:13  <ikaros> yea there must something different be wrong
2011-12-08T16:26:25  <ikaros> this is strange
2011-12-08T16:26:35  <McLeopold> last contest waiting till the end was bad, because of the engine, now you can and it's ok
2011-12-08T16:26:54  <besh> I don't understand why the first one is better ? Infact I deliberately changed to the second one in the hope it will be faster , silly me :(
2011-12-08T16:27:27  *** Anilm3 <Anilm3!~anil@4.Red-79-146-241.dynamicIP.rima-tde.net> has joined #aichallenge
2011-12-08T16:27:31  *** nplus <nplus!~chatzilla@unaffiliated/simont> has joined #aichallenge
2011-12-08T16:27:46  <BenJackson> besh: the engine's main loop is:  while sleep 10ms: for all alive bots read up to 100 lines
2011-12-08T16:28:13  <BenJackson> McLeopold: So I thought I had a clever idea to get better CPU utilization on the workers and then I saw it was already implemented
2011-12-08T16:28:22  *** heinrich5991 <heinrich5991!~hein5991@unaffiliated/heinrich5991> has quit IRC (Excess Flood)
2011-12-08T16:28:23  <Anilm3> can some one explain me how can make my antes go to objectives like lazarant on this match? http://aichallenge.org/visualizer.php?game=173325&user=589
2011-12-08T16:28:24  <McLeopold> the 2 at a time?
2011-12-08T16:28:24  <BenJackson> is the live setup actually using --serial or some variant?
2011-12-08T16:28:47  <BenJackson> in the config template I see 'serial':2 but --serial 2 is not a valid cmd line (it's just a boolean)
2011-12-08T16:28:48  <McLeopold> serial 2
2011-12-08T16:29:00  <McLeopold> serial is only for the engine, not playgame
2011-12-08T16:29:08  <McLeopold> the official contest doesn't use playgame
2011-12-08T16:29:13  <McLeopold> that is local play only
2011-12-08T16:29:14  <BenJackson> playgame does take it, ah
2011-12-08T16:29:20  <BenJackson> ok, cool
2011-12-08T16:29:32  <BenJackson> so maybe having CPU time controls vs wall clock is not that big a deal
2011-12-08T16:29:38  <McLeopold> I could added... or you could do a pull request
2011-12-08T16:29:39  <BenJackson> since it's not up to 9 bots waiting on the slowest
2011-12-08T16:29:54  <McLeopold> right
2011-12-08T16:29:56  <BenJackson> McLeopold: no, I don't need it locally, I was just brainstorming about the workers
2011-12-08T16:30:06  <BenJackson> and then saw the code was there but didn't understand the playgame/engine distinction
2011-12-08T16:30:17  <Fluxid> http://ants.fluxid.pl/replay.19470 great method, have 260 ants in hill so you're safe for 260 turns
2011-12-08T16:30:26  <BenJackson> Fluxid: ChrisH does that
2011-12-08T16:30:32  <BenJackson> I just implemented hive count
2011-12-08T16:30:32  <BenJackson> but I'
2011-12-08T16:30:36  <BenJackson> m not using for anything yet
2011-12-08T16:30:54  <rwest> I used to do that on accident
2011-12-08T16:30:56  <Fluxid> okay okay, lemme quickly hack constant "turntime 500"
2011-12-08T16:31:00  <besh> bj: why doesn't it read everything in the socket buffer ?
2011-12-08T16:31:05  <rwest> my record is 960 in the hill
2011-12-08T16:31:08  <rwest> I think I was safe
2011-12-08T16:31:15  <BenJackson> besh: unknown
2011-12-08T16:31:41  <BenJackson> you're safe as soon as hive() > (state.turns - state.turn)
2011-12-08T16:31:48  <BenJackson> as long as your logic doesn't move ants off your hill
2011-12-08T16:32:08  <BenJackson> BUT your hills are only worth 1 point so you can probably still score higher if you bring more ants to the party
2011-12-08T16:32:13  <BenJackson> once your hill is surrounded you won't be able to
2011-12-08T16:32:14  <Fluxid> 12 games running on my tcp
2011-12-08T16:32:22  <Fluxid> can i kill them?
2011-12-08T16:32:28  <besh> fluxid: thumbs up for turntime 500
2011-12-08T16:32:48  *** Redgis <Redgis!~Nex@2a01:e34:ee3b:aa40:ccb:8a1d:94b0:4c6> has quit IRC (Quit: ... mains libres)
2011-12-08T16:33:04  <BenJackson> fluxid: statistically speaking most people are not winning those 12 games
2011-12-08T16:33:08  <BenJackson> so they all vote yet :)
2011-12-08T16:33:08  <BenJackson> s
2011-12-08T16:33:14  *** heinrich5991 <heinrich5991!~hein5991@unaffiliated/heinrich5991> has joined #aichallenge
2011-12-08T16:33:16  <McLeopold> yets?
2011-12-08T16:33:24  <BenJackson> nyet
2011-12-08T16:33:32  <McLeopold> :)
2011-12-08T16:33:55  <Fluxid> okay then!
2011-12-08T16:34:00  <Fluxid> KILL ALL THE GAMES
2011-12-08T16:34:07  <McLeopold> Fluxid: so, how about that turn time?
2011-12-08T16:34:36  <Fluxid> McLeopold: now it should send turntime 500
2011-12-08T16:34:40  <Fluxid> someone check
2011-12-08T16:34:48  <thestinger> it's fast now
2011-12-08T16:34:50  <thestinger> :)
2011-12-08T16:35:10  <thestinger> my cpu load is way up which is a good sign
2011-12-08T16:35:12  <BenJackson> mine won't be restarted until I get home
2011-12-08T16:35:34  <besh> yes it seems to go faster but i have to wait till a new game starts ..
2011-12-08T16:36:32  <besh> no it is still 5000
2011-12-08T16:37:40  <Fluxid> besh: you mean you get 5000 or you see 5000 in game replay on site
2011-12-08T16:38:00  <Fluxid> because it replaces turntime when sending to socket with 500
2011-12-08T16:38:05  <Fluxid> not in engine itself
2011-12-08T16:38:16  <Fluxid> at least it should
2011-12-08T16:38:20  <ikaros> wasnt there a problem with the timer from the c++ starter pack?
2011-12-08T16:38:30  <thestinger> ikaros: I think it was broken
2011-12-08T16:38:40  <ikaros> ok.. best ill look into that first :)
2011-12-08T16:39:06  <thestinger> the best way to do it is saving a timeval struct, and calculating a delta based on a new one
2011-12-08T16:39:31  <thestinger> instead of trying to turn unix time seconds into milli/microseconds
2011-12-08T16:39:59  <besh> fluxid: I see 5000 for both loadtime and turntime but the games to go much faster..
2011-12-08T16:40:17  <Fluxid> i'll check myself with my bot then
2011-12-08T16:40:43  <Fluxid> it's 500
2011-12-08T16:40:46  <g0llum> fluxid, what did you do to achieve the 500 output t the bots ?
2011-12-08T16:40:58  <Fluxid> g0llum: i'll commit and push in a second
2011-12-08T16:41:03  <thestinger> ikaros: http://paste.pocoo.org/show/z5jGMzQw7jxcJkOOYa4B/ I just turned it into that (you can ignore the silly anonymous namespace)
2011-12-08T16:41:30  *** kire <kire!~kire@78-22-125-201.access.telenet.be> has joined #aichallenge
2011-12-08T16:41:33  <ikaros> thx i was just reading up clock_gettime .. no idea whats the better choice
2011-12-08T16:41:37  <LouisMartin> greeting gentlemen  :)
2011-12-08T16:41:56  <thestinger> ikaros: clock_gettime is the modern way (gettimeofday is deprecated according to the POSIX:2008 docs)
2011-12-08T16:42:02  *** olexs <olexs!~Leviathan@drms-590c5f5b.pool.mediaWays.net> has joined #aichallenge
2011-12-08T16:42:05  <thestinger> but I think you need -lrt passed to gcc to use the clock_ stuff
2011-12-08T16:42:11  <Fluxid> g0llum: https://github.com/fluxid/ants-tcp/commit/4cdd4c7c293c46094271020b4b7f93858c2e62e3 here you are
2011-12-08T16:42:11  <ikaros> ohh
2011-12-08T16:42:19  <ikaros> ok thx i will try your version first
2011-12-08T16:42:57  <thestinger> the new clock_ stuff lets you use CLOCK_MONOTONIC (based on an arbitrary time) instead of wall clock time, which can be changed by ntpd, etc.
2011-12-08T16:43:17  <LouisMartin> as I'm about to begin the battle code for my post-nervous breakdown v2 bot, I was wondering if some of you would be kind enough to give me a few advices and tips? share some horror stories of what to avoid?
2011-12-08T16:43:26  <thestinger> and I think it's more accurate (the struct uses nsec, not usec :P) and you can get the resolution with clock_getres
2011-12-08T16:43:32  <thestinger> gettimeofday() is fine though
2011-12-08T16:43:46  *** kire <kire!~kire@78-22-125-201.access.telenet.be> has quit IRC (Remote host closed the connection)
2011-12-08T16:44:25  <ikaros> gah still timouts.. so it wasnt the timers fault.. wtf
2011-12-08T16:45:35  <besh> fluxid: I am getting 5000 for the third time even after I restarted ? Replay also says 5000
2011-12-08T16:45:41  <LouisMartin> has anyone shared some scenarios maps to practice battle / battle aspect and such?   is there a map editor or any precious tool I should know about?
2011-12-08T16:46:43  <McLeopold> LouisMartin: notepad...
2011-12-08T16:47:13  <thestinger> I think there's an emacs mode for drawing ascii art...
2011-12-08T16:47:20  <BenJackson> to be 100% accurate:  clock_gettime CLOCK_MONOTONIC's *rate* is affected by NTP, but not its offset
2011-12-08T16:47:40  <BenJackson> but gettimeofday and clock_gettime use the same underlying mechanism
2011-12-08T16:47:45  <thestinger> ah
2011-12-08T16:47:48  <BenJackson> and linux actually stores the wall clock time and translates to monotonic
2011-12-08T16:47:56  <BenJackson> so gettimeofday is microscopically faster
2011-12-08T16:48:28  <McLeopold> oh wow, you drop all games when restarting?
2011-12-08T16:48:48  <LouisMartin> McLeopold:  Yea, I was thinking of TheDraw (http://en.wikipedia.org/wiki/TheDraw)
2011-12-08T16:49:17  <thestinger> http://www.lysator.liu.se/~tab/artist/
2011-12-08T16:49:41  <thestinger> anyway I'm just doing it by hand in vim :P
2011-12-08T16:50:37  <BenJackson> yeah, I have a template arena
2011-12-08T16:50:43  <BenJackson> and I modify it in vim
2011-12-08T16:50:54  <LouisMartin> thestinger: I like it
2011-12-08T16:51:36  <g0llum> fluxid, still got turntime 5000fluxid, it's not the bot, that writes turntime, but the angine/ants.py(1542) i think, you mixed up bot read/write
2011-12-08T16:52:13  <Fluxid> lemme check once more
2011-12-08T16:53:09  <LouisMartin> so, no maps/bots to practice battle code has been shared?
2011-12-08T16:53:30  <McLeopold> a few on paste
2011-12-08T16:53:42  <Fluxid> hm, you're right
2011-12-08T16:55:05  <thestinger> in vim, I just use block visual with ctrl-V and then r% to change to walls r. to change to land, etc.
2011-12-08T16:55:20  <thestinger> the emacs artist mode thing is kind of neat though
2011-12-08T16:55:28  *** _Lone_Wolf_ <_Lone_Wolf_!LW@bl10-78-112.dsl.telepac.pt> has joined #aichallenge
2011-12-08T16:55:34  *** _Lone_Wolf_ is now known as flowenol
2011-12-08T16:55:53  <LouisMartin> McLeopold: thanks
2011-12-08T16:55:57  <BenJackson> thestinger: you'd like aacircuit
2011-12-08T16:57:19  <thestinger> lol, my bot loses to an ancient pre-combat version on small maps
2011-12-08T16:57:25  <thestinger> 15% win rate
2011-12-08T16:57:48  <thestinger> on tiny random walks
2011-12-08T16:58:20  <besh> do most people use the same name for different versions when playing on tcp ?
2011-12-08T16:58:43  <flowenol> besh idk but i do
2011-12-08T16:58:47  <thestinger> besh: I use strcat and strcat_next for my stable version and master branch
2011-12-08T16:58:48  <McLeopold> just one http://ants.fluxid.pl/player/McLeopold_sucks
2011-12-08T16:58:57  <thestinger> the random ones with suffixes are just branches I want to test
2011-12-08T16:59:05  <besh> thestinger: yeah we know what you do :)
2011-12-08T16:59:10  <thestinger> :P
2011-12-08T16:59:25  <thestinger> adding a version number and starting at a high sigma again just seems like a waste
2011-12-08T16:59:30  <besh> but i think what I do is bad practise as it could affect ratings
2011-12-08T17:01:40  <BenJackson> on fluxid you can get 200 games a day
2011-12-08T17:01:46  <BenJackson> restarting hardly matters
2011-12-08T17:02:45  <Fluxid> one more chance
2011-12-08T17:03:00  <BenJackson> with this fix mayb 400 games a day!
2011-12-08T17:03:02  <thestinger> if all else fails, hardcode an iptables rule to edit the packet P
2011-12-08T17:03:04  <thestinger> :P *
2011-12-08T17:03:09  <Fluxid> OKAY
2011-12-08T17:03:14  <Fluxid> now it should be good
2011-12-08T17:03:37  <Fluxid> fucking python 2.6 on the server
2011-12-08T17:03:38  <BenJackson> to make fluxid more like the main site I'm going to connect with 3500 starter kits
2011-12-08T17:03:41  <besh> yes now it works
2011-12-08T17:03:49  <Fluxid> good
2011-12-08T17:04:00  <Fluxid> sorry for the restarts and enjoyyyyyyyy~~~
2011-12-08T17:04:08  <besh> thanks: Now I should have 10x more games that I don't know what to do with :)
2011-12-08T17:04:23  <Anilm3> do your ants give priority to fighting or gathering food?
2011-12-08T17:04:27  <BenJackson> besh: I did a few statistical analyses of all my fluxid games
2011-12-08T17:04:31  <BenJackson> looking at map sizes I was bad at
2011-12-08T17:04:33  <BenJackson> particular maps, etc
2011-12-08T17:04:37  <BenJackson> then using that to pick replays
2011-12-08T17:04:47  <thestinger> http://ants.fluxid.pl/player/hedonismbot this is mine btw :P
2011-12-08T17:04:54  <thestinger> ripped out 50% of the logic and made it only care about food
2011-12-08T17:05:07  <thestinger> seems to be an improvement...
2011-12-08T17:05:08  <BenJackson> gold-plated grape-eating chaise lounge robot?
2011-12-08T17:05:10  *** avdg <avdg!~avdg@78-22-168-200.access.telenet.be> has quit IRC (Ping timeout: 248 seconds)
2011-12-08T17:05:23  *** mviel__ <mviel__!~mviel@213.163.164.204> has quit IRC (Ping timeout: 252 seconds)
2011-12-08T17:05:42  <Fluxid> oh btw, i updated the maps
2011-12-08T17:05:51  <BenJackson> http://images.wikia.com/en.futurama/images/1/18/Hedonismbot.jpg
2011-12-08T17:05:58  <thestinger> BenJackson: yep
2011-12-08T17:05:59  <besh> bj: that is interesting. I know my bot does bad at open maps but it is good to have a statistical check.
2011-12-08T17:06:06  <BenJackson> Fluxid: did you have to buy new disks to hold them all?
2011-12-08T17:06:14  <BenJackson> cell_maze_05p_9249249298.map
2011-12-08T17:06:51  <Fluxid> uh, no, not really. i just copied them from aichallenge repo
2011-12-08T17:07:05  <BenJackson> is that up to date?
2011-12-08T17:07:09  <BenJackson> I did that recently and was still missing maps
2011-12-08T17:07:13  <Fluxid> http://ants.fluxid.pl/maps i think so
2011-12-08T17:07:33  *** mviel__ <mviel__!~mviel@213.163.164.204> has joined #aichallenge
2011-12-08T17:07:54  <Fluxid> lots of... maps
2011-12-08T17:07:56  <thestinger> aw crap, there are a bunch of random walks now
2011-12-08T17:08:06  <BenJackson> 300 maps, wow
2011-12-08T17:08:14  <Fluxid> you'll need to play for two days to make sure you played all of them...
2011-12-08T17:08:26  <thestinger> BenJackson: all the random walks are basically the same though :P
2011-12-08T17:08:30  <BenJackson> yeah that seems to be a big disparity for the finals
2011-12-08T17:08:38  <Fluxid> now noone will will even try to hardcode map-dependent behaviour
2011-12-08T17:08:39  <BenJackson> you and some other bot could play entirely disjoint sets of maps
2011-12-08T17:08:55  <thestinger> well, this sucks
2011-12-08T17:09:01  <BenJackson> thestinger: I found a unique feature of 1
2011-12-08T17:09:13  <BenJackson> an ant could hide in a niche such that only one enemy ant could threaten it
2011-12-08T17:09:20  <BenJackson> my code kept trying to send more ants to finish it off
2011-12-08T17:09:26  <BenJackson> but none could get close enough
2011-12-08T17:09:32  <thestinger> this map I'm testing on is silly, my bots raise each other's hill at the same turn
2011-12-08T17:09:34  <BenJackson> so it boxed that one ant in with 10+ ants :(
2011-12-08T17:09:38  <thestinger> almost every game
2011-12-08T17:13:44  *** iglo <iglo!~xy@xdslaf024.osnanet.de> has joined #aichallenge
2011-12-08T17:13:45  <BenJackson> in explore tests it turns out my bot with combat but no goals (like attacking hills) tends to kill all the holdbots anyway
2011-12-08T17:13:52  <BenJackson> I'm not entirely sure why
2011-12-08T17:14:06  <BenJackson> I think killing the holdbots must get me close enough to the hills that I accdientally step on them
2011-12-08T17:14:19  <thestinger> lol
2011-12-08T17:16:26  <BenJackson> part of the reason I added --forever to the game
2011-12-08T17:16:31  <BenJackson> to shut off the endgame conditions
2011-12-08T17:18:30  <thestinger> http://ompldr.org/vYm5oaw all my tests end like this...
2011-12-08T17:18:49  <Fluxid> lol
2011-12-08T17:19:01  <Fluxid> perfect tie
2011-12-08T17:19:33  *** mleise <mleise!~marco@dslb-088-070-148-091.pools.arcor-ip.net> has quit IRC (Ping timeout: 248 seconds)
2011-12-08T17:20:38  <thestinger> I think I fixed a bug
2011-12-08T17:20:49  <thestinger> now my new version seems to win most of the time
2011-12-08T17:22:04  <BenJackson> you know I keep meaning to figure out why I can run n versions of my bot on the same map
2011-12-08T17:22:09  <Fluxid> like it didn't before
2011-12-08T17:22:13  <BenJackson> without any rotations in the symmetry
2011-12-08T17:22:18  <BenJackson> yet they won't do identical things
2011-12-08T17:23:02  <thestinger> your direction choices are probably biased
2011-12-08T17:23:08  <thestinger> mine are for sure
2011-12-08T17:23:29  *** nplus <nplus!~chatzilla@unaffiliated/simont> has quit IRC (Read error: Connection reset by peer)
2011-12-08T17:24:22  <thestinger> what's the max name length on tcp?
2011-12-08T17:24:59  <thestinger> oh, apparently there is none
2011-12-08T17:25:59  <Fluxid> inb4 exploit
2011-12-08T17:26:02  <g0llum> depends on your will to cheat, since it's in the client ;)
2011-12-08T17:28:28  <thestinger> I need to redo movement again, my ants think they are bees atm
2011-12-08T17:28:45  <thestinger> I like how some bots move around like blobs of mercury
2011-12-08T17:30:38  *** tobym <tobym!~tobym@pool-71-183-248-125.nycmny.fios.verizon.net> has quit IRC (Remote host closed the connection)
2011-12-08T17:32:02  *** tobym <tobym!~tobym@pool-71-183-248-125.nycmny.fios.verizon.net> has joined #aichallenge
2011-12-08T17:32:08  <besh> fluxid: I tried your idea about cost of path through ants for A* path finding, and it helps.
2011-12-08T17:32:51  <besh> I increment the cost by +2 when the path is through an ant. Maybe too aggresive ?
2011-12-08T17:33:17  <besh> I am wondering if I can manipulate the cost to include "unseen tile" information.
2011-12-08T17:33:50  <thestinger> besh: I do +3 when going through an ant instead of the usual +1
2011-12-08T17:34:09  <besh> Aha so we do similar things then.
2011-12-08T17:34:11  <thestinger> well, I've tried like a hundred variations
2011-12-08T17:34:40  <besh> It helps to avoid path through pile of your ants like at choke points.
2011-12-08T17:35:13  <thestinger> yeah, they take alternate routes if it's not too much further
2011-12-08T17:35:32  <thestinger> I need a better way of doing it though
2011-12-08T17:36:34  <besh> I am thinking to use it for maximizing exploration as well. To prefer paths through less visited tiles more but getting the right weights may be difficult.
2011-12-08T17:39:01  <besh> My A* paths are not really that optimal anyway. I use  weighted A* with a factor of 2
2011-12-08T17:39:57  *** Paradoxial <Paradoxial!~Paradoxia@pool-108-28-22-94.washdc.fios.verizon.net> has joined #aichallenge
2011-12-08T17:42:33  <besh> thestinger: someone has the same exact confusions as I have about the fighting[][] array http://forums.aichallenge.org/viewtopic.php?f=24&t=2083
2011-12-08T17:43:24  *** Jak_o_Shadows <Jak_o_Shadows!~hayden@CPE-144-136-214-52.sa.bigpond.net.au> has joined #aichallenge
2011-12-08T17:47:04  *** nplus <nplus!~chatzilla@unaffiliated/simont> has joined #aichallenge
2011-12-08T17:47:38  <thestinger> antimatroid: heh, I actually dealt with your example (with the two ants not being able to go to the same place) already
2011-12-08T17:48:14  <thestinger> because I put all possible locations into an unordered_set, then paint the info around them
2011-12-08T17:48:48  <thestinger> oh nvm, I need to tweak it a tiny bit to deal with it properly
2011-12-08T17:50:35  <besh> If there was no water, my code tries to order moves to get all ants closer to the target.
2011-12-08T17:50:56  <besh> but that water tile forces them to move to the same location for his example.
2011-12-08T17:53:25  *** bluegaspode <bluegaspode!55b223f9@gateway/web/freenode/ip.85.178.35.249> has quit IRC (Quit: Page closed)
2011-12-08T17:54:04  *** ikaros <ikaros!~ikaros@dslb-084-059-073-057.pools.arcor-ip.net> has quit IRC (Quit: Ex-Chat)
2011-12-08T17:54:10  <LouisMartin> hmm, what is --scenario used for?  playgame.py usage info doesn't tell
2011-12-08T17:55:04  <BenJackson> to feed in a map with preexisting food and ants
2011-12-08T17:55:06  <BenJackson> for testing
2011-12-08T17:55:12  <McLeopold> and ignoring symmetry
2011-12-08T17:55:14  <BenJackson> you will want --food none too in most cases
2011-12-08T17:55:21  <McLeopold> or --food random
2011-12-08T17:56:38  <LouisMartin> ah! that's what I thought. the engine doesn't like my idea of a scenario though..  "got 43 rows, expected none"
2011-12-08T17:58:57  <LouisMartin>   File "C:\Local Workspace\AIC Ants\tools\ants.py", line 271, in parse_map     % (height, row)) Exception: ('map', 'Incorrect number of rows.  Expected None, got 43')
2011-12-08T17:59:27  *** ikaros <ikaros!~ikaros@dslb-084-059-073-057.pools.arcor-ip.net> has joined #aichallenge
2011-12-08T17:59:28  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has quit IRC (Ping timeout: 240 seconds)
2011-12-08T18:00:18  <LouisMartin> I'm an idiot, but why?  (in the context of this error, plz...)
2011-12-08T18:01:25  <BenJackson> bad map?
2011-12-08T18:01:44  <g0llum> LouisMartin, now the whole cmdline, please
2011-12-08T18:01:46  <BenJackson> put your map on paste.aichallenge
2011-12-08T18:02:16  <LouisMartin> perhaps.. I'll check again
2011-12-08T18:03:36  <LouisMartin> I'll clean it first as it seems too loaded for nothing, g0llum. I'm recycling the same cmd line over and over and it probably ain't good for scenario map.
2011-12-08T18:03:57  <LouisMartin> also, as Ben suggest, my map is probably just bad
2011-12-08T18:05:24  <g0llum> maybe, you just forgot to say rows 20 and the like in the map
2011-12-08T18:07:59  *** iglo <iglo!~xy@xdslaf024.osnanet.de> has quit IRC (Remote host closed the connection)
2011-12-08T18:08:14  <LouisMartin> I'll paste it
2011-12-08T18:08:31  <LouisMartin> I don't get it since the rows and cols declared in the map are okays
2011-12-08T18:09:59  <LouisMartin> http://paste.aichallenge.org/v9IY1/
2011-12-08T18:10:40  *** sigh <sigh!~sigh@ppp114-153.static.internode.on.net> has joined #aichallenge
2011-12-08T18:11:27  <LouisMartin> the command line is "python "%~dp0playgame.py" -So --engine_seed 62 --player_seed 59 --end_wait=0.25 --verbose --log_dir game_logs --loadtime 36000 --turntime 2000 --turns 36000 --  viewradius2 77 --food_rate 5 11 --food_turn 19 37 --food_start 75 175 --food_visible 3 5 --scenario --food none --map_file "%~dp0maps\cases\battledummy.map" %*   "python ""%~dp0sample_bots\python\LeftyBot.py""" """%~dp0mybots\realsmurfs.exe""" | j
2011-12-08T18:11:31  <g0llum> i've seen that map before!
2011-12-08T18:11:41  <thestinger> I made it to test my early combat :P
2011-12-08T18:11:47  <thestinger> it's just the tutorial map with some ants
2011-12-08T18:12:00  <LouisMartin> and yes, it needs a redesign to say the least  lol
2011-12-08T18:12:09  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has joined #aichallenge
2011-12-08T18:12:15  <BenJackson> I found it somewhat useful to see how many ants my orange team could survive with
2011-12-08T18:12:44  <LouisMartin> thestinger :)   yes, you're the one who gave it to me if I remember well.
2011-12-08T18:12:57  <thestinger> http://ompldr.org/vYm5pcw red == die, blue == trade
2011-12-08T18:13:09  <BenJackson> haha I thought this new code would be slower
2011-12-08T18:13:10  <LouisMartin> or maybe it just got viral
2011-12-08T18:13:12  <BenJackson> it's 10% faster
2011-12-08T18:16:12  <LouisMartin> so this map is well known and tested; how could it fail on my side?
2011-12-08T18:16:40  <LouisMartin> the starter pack "tutorial1.map" runs perfectly with the same cmd line
2011-12-08T18:16:46  *** oddmunds <oddmunds!~oddmunds@s21-04214.dsl.no.powertech.net> has left #aichallenge
2011-12-08T18:17:51  <g0llum> the empty line ? why not just diff those ?
2011-12-08T18:20:03  <LouisMartin> I don't understand. which empty line?  in the map?  those are okay by the format, but I'll try
2011-12-08T18:21:27  <g0llum> empty line between players 2 and actual map data, other maps don't have that
2011-12-08T18:21:36  <LouisMartin> btw, what would be the parameter  -So  ?   I blindly copy/pasted it for a while without knowing. It's not in the playgame.py usage
2011-12-08T18:22:28  <BenJackson> LouisMartin: bad line endings?
2011-12-08T18:22:45  <LouisMartin> tried it, but it doesn't work
2011-12-08T18:23:50  <LouisMartin> Ben: line endings?    I thought notepad was okay?
2011-12-08T18:24:10  <LouisMartin> cr/lf  versus just lf  ?
2011-12-08T18:25:07  <LouisMartin> I actually used Visual Studio to save it, not notepad
2011-12-08T18:28:45  <LouisMartin> ...and that is it: saving maps under common windows editors adds the <Cr> and playgame.py dislike those. I opened and <saved as> a perfectly working map and now it's ruined.
2011-12-08T18:29:18  <LouisMartin> thanks BenJackson and g0llum. mystery solved.
2011-12-08T18:32:22  <LouisMartin> gotta love "not invented here syndrome":  Windows has <Cr/Lf>, Linux <Lf> and Apple/Mac <Cr>
2011-12-08T18:33:46  <thestinger> OSX does it properly now iirc
2011-12-08T18:33:49  <thestinger> some applications may not
2011-12-08T18:33:53  <thestinger> \n is the one true way
2011-12-08T18:33:56  <thestinger> who cares what windows does :P
2011-12-08T18:34:00  <g0llum> yea, almost 2012, and we still splitting hairs on that
2011-12-08T18:34:09  <thestinger> windows also doesn't use UTF-8 everywhere
2011-12-08T18:34:25  <thestinger> and I think notepad/wordpad have broken unicode
2011-12-08T18:34:36  <BenJackson> you could probably fix the engine to load crlf maps pretty easily
2011-12-08T18:34:39  <thestinger> because they use UCS-2, not UTF-16
2011-12-08T18:35:14  <Antimony> Wow, apparently my bot doesn't work on square maps
2011-12-08T18:35:27  *** epicmonkey <epicmonkey!~epicmonke@178.130.5.103> has quit IRC (Ping timeout: 240 seconds)
2011-12-08T18:35:59  <thestinger> wikipedia says win8 actually uses LF (\n) line endings
2011-12-08T18:36:00  <LouisMartin> BenJackson: indeed an option
2011-12-08T18:36:06  *** avdg <avdg!~avdg@78-22-168-200.access.telenet.be> has joined #aichallenge
2011-12-08T18:36:06  <thestinger> might not be true
2011-12-08T18:38:30  <a1k0n> ÿþhi i'm a windows unicode text file
2011-12-08T18:38:39  <a1k0n> http://en.wikipedia.org/wiki/Byte_order_mark
2011-12-08T18:40:31  *** tobel <tobel!~user@pD9E8A32A.dip.t-dialin.net> has quit IRC (Quit: ERC Version 5.3 (IRC client for Emacs))
2011-12-08T18:40:39  *** AntDroid <AntDroid!70c6400f@gateway/web/freenode/ip.112.198.64.15> has joined #aichallenge
2011-12-08T18:40:58  *** cyphase <cyphase!~cyphase@unaffiliated/cyphase> has quit IRC (Ping timeout: 268 seconds)
2011-12-08T18:42:07  <LouisMartin> cause there's one thing Microsoft loves more than fair competition; standards
2011-12-08T18:42:49  *** Conorach <Conorach!~cono@90.170.78.100> has quit IRC (Ping timeout: 268 seconds)
2011-12-08T18:44:03  *** iglo <iglo!~xy@xdslaf024.osnanet.de> has joined #aichallenge
2011-12-08T18:44:12  <thestinger> http://harmful.cat-v.org/standards/
2011-12-08T18:45:29  <LouisMartin> thestinger: you seem well versed in text encoding, any idea what I should feed visual studio when saving?  it asks for encoding and page ending
2011-12-08T18:46:34  <thestinger> page ending?
2011-12-08T18:47:01  <thestinger> utf-8 is the only sane encoding
2011-12-08T18:47:16  <thestinger> I think most text editors add an extra newline at the end
2011-12-08T18:47:43  *** mviel_ <mviel_!~mviel@236.227.73.86.rev.sfr.net> has joined #aichallenge
2011-12-08T18:48:24  <LouisMartin> 'Mediocrity finds safety in standardization.'       XML
2011-12-08T18:48:36  <thestinger> there's also a great page on XML on there
2011-12-08T18:48:47  <thestinger> http://harmful.cat-v.org/software/xml/
2011-12-08T18:49:01  <LouisMartin> extra newline on the end?
2011-12-08T18:49:12  <thestinger> it probably puts it there
2011-12-08T18:49:19  <thestinger> you just can't see it in the editor
2011-12-08T18:50:17  <thestinger> Java is a DSL to transform big XML documents into long exception stack traces. -- Scott Bellware
2011-12-08T18:50:22  <thestinger> best quote ever :P
2011-12-08T18:50:24  <LouisMartin> oh god, site is quite therapeutic
2011-12-08T18:51:19  <thestinger> I think http://harmful.cat-v.org/ is done by uriel (one of the suckless.org people)
2011-12-08T18:51:23  *** mviel__ <mviel__!~mviel@213.163.164.204> has quit IRC (Ping timeout: 252 seconds)
2011-12-08T18:51:24  <BenJackson> ""XML combines the efficiency of text files with the readability of binary files" -- unknown
2011-12-08T18:51:27  <BenJackson> nice
2011-12-08T18:51:35  <LouisMartin> my stomach hurts good :D    truth win
2011-12-08T18:51:36  *** bluegaspode <bluegaspode!55b223f9@gateway/web/freenode/ip.85.178.35.249> has joined #aichallenge
2011-12-08T18:52:02  <McLeopold> LouisMartin: -So is the same as -S and -o
2011-12-08T18:52:19  <McLeopold> stream output and output to stdout
2011-12-08T18:52:48  <McLeopold> it is needed if playing a game live with the java visualizer
2011-12-08T18:52:53  <thestinger> anyway uriel is a rob pike fanboy so there's lots of *nix, plan9, Go, C, etc. stuff on cat-v.org
2011-12-08T18:53:02  *** cyphase <cyphase!~cyphase@unaffiliated/cyphase> has joined #aichallenge
2011-12-08T18:53:14  <g0llum> go_ants_go!
2011-12-08T18:53:44  <LouisMartin> ah okay. So I guess it I can leave it there without paranoia.
2011-12-08T18:54:11  <LouisMartin> bookmarked
2011-12-08T18:54:11  *** bluegaspode <bluegaspode!55b223f9@gateway/web/freenode/ip.85.178.35.249> has quit IRC (Client Quit)
2011-12-08T18:59:37  <LouisMartin> I can't believe this. VS has 6 options for line ending and none - when combien with UTF-8 (with or without signature, whatever this means) saves the map correctly
2011-12-08T19:00:47  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has quit IRC (Ping timeout: 240 seconds)
2011-12-08T19:01:28  *** bleh <bleh!a8698c48@gateway/web/freenode/ip.168.105.140.72> has joined #aichallenge
2011-12-08T19:01:51  *** fourmidable <fourmidable!adb2c167@gateway/web/freenode/ip.173.178.193.103> has quit IRC (Quit: Page closed)
2011-12-08T19:01:54  *** bleh is now known as Guest74161
2011-12-08T19:04:00  <g0llum> LouisMartin, you won't guess how many bullshit you can pack into a ~200mb ide ;(   (still the best c++ one i've seen so far)
2011-12-08T19:04:40  <BenJackson> haha, my play_one_game is modified to automatically supply the right number of opponents
2011-12-08T19:04:51  <BenJackson> I did a run without my bot on the command line
2011-12-08T19:04:54  <BenJackson> and didn't notice
2011-12-08T19:05:03  <BenJackson> so I kept running my various test opponents against teach other
2011-12-08T19:08:45  <LouisMartin> g0llum: exact.  this is what you deserve for insisting on being MS little b'tch after so many years of awesome open source nix dev efforts available instantly, open and free. I totally deserve this. fml
2011-12-08T19:09:16  <LouisMartin> BenJackson: Who won? :}
2011-12-08T19:09:41  <besh> "Next game could take a while" Almost a day now
2011-12-08T19:10:03  <g0llum> LouisMartin, still stick to it, just don't insist topen 'map files' with that
2011-12-08T19:10:42  *** Guest74161 <Guest74161!a8698c48@gateway/web/freenode/ip.168.105.140.72> has quit IRC (Quit: Page closed)
2011-12-08T19:11:43  <LouisMartin> g0llum: until the end of the contest, sure.. no plan on getting excited on learning python in 10 days.
2011-12-08T19:12:12  <_flag> Is it just me or has the game rate decreased?
2011-12-08T19:12:18  <LouisMartin> also: I've succeeded with "US-ASCII codepage 20127" and "Unix LF" line ending
2011-12-08T19:12:40  <_flag> There appears to only be one worker running, so that might be why
2011-12-08T19:12:44  <BenJackson> LouisMartin: actually the "best" one (the highest submitted version)
2011-12-08T19:12:52  <besh> _flag: it has decreased signifcantly. Yesterday was 0.0 games / minute for me
2011-12-08T19:12:55  <g0llum> besh, we're all waiting for raspberryPi to deliver, -- the next contest will be run from 24 worker running out of a shoebox , poered by solar energy.
2011-12-08T19:13:09  <BenJackson> I played 0 games overa bout 25 hours and then 5 or so in 10 minutes
2011-12-08T19:13:25  <_flag> McLeopold: Not sure if you're aware, but 5 out of 6 workers are down
2011-12-08T19:13:31  <_flag> janzert: ^
2011-12-08T19:13:45  <Anilm3> I can't keep watching my bot suffer on Fluxid :(
2011-12-08T19:14:57  <LouisMartin> _flag: I've seen something to the effect that the servers were currently costly and so they release a few to keep on budget.
2011-12-08T19:15:52  <LouisMartin> It's currently 20x more costly than commonly. They're keeping the server budget for running the finals.
2011-12-08T19:16:50  <McLeopold> _flag: yep, but I don't have access to revive them, that's janzert only
2011-12-08T19:18:39  <LouisMartin> McLeopold: So the low rate of game played isn't related to the lowered number of servers as much as it's a technical problem (down workers) then?
2011-12-08T19:20:27  *** g0llum <g0llum!~g0llum@82.113.99.35> has quit IRC (Read error: Connection reset by peer)
2011-12-08T19:20:28  *** olexs <olexs!~Leviathan@drms-590c5f5b.pool.mediaWays.net> has quit IRC (Quit: Leaving.)
2011-12-08T19:23:29  <McLeopold> I don't see a price increase, so they should be running
2011-12-08T19:23:51  <McLeopold> http://cloudexchange.org/charts/us-east-1.linux.m2.xlarge.html
2011-12-08T19:26:34  <ronchaine> My ants!  They do something!
2011-12-08T19:26:46  *** iglo <iglo!~xy@xdslaf024.osnanet.de> has quit IRC (Remote host closed the connection)
2011-12-08T19:27:52  <BenJackson> wow win prices are a lot higher
2011-12-08T19:29:05  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has joined #aichallenge
2011-12-08T19:30:31  <BenJackson> no wonder replay.html is taking a while to load across the internet... it's 50M with visualization
2011-12-08T19:32:16  <thestinger> lol
2011-12-08T19:32:18  <thestinger> yeah
2011-12-08T19:33:50  *** Migi32 <Migi32!~migi@178-117-102-152.access.telenet.be> has joined #aichallenge
2011-12-08T19:37:49  <Antimony> ug my symmetry detection is so slow that the bot has already found hills by the time it's done
2011-12-08T19:39:04  <McLeopold> http://paste.aichallenge.org/37Zbn/
2011-12-08T19:39:34  <McLeopold> I'd just like to say the visualization overlay is awesome
2011-12-08T19:42:21  <besh> antimony: be glad you have symmetry detection at all.
2011-12-08T19:42:30  <besh> i have none :(
2011-12-08T19:43:05  <Anilm3> does symmetry detection really worth it?
2011-12-08T19:43:32  <McLeopold> I guess it depends on how fast it is, in turns
2011-12-08T19:43:32  <besh> I think it helps if you want to attack hills early
2011-12-08T19:43:40  <McLeopold> it would get you hill locations
2011-12-08T19:44:06  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has quit IRC (Ping timeout: 248 seconds)
2011-12-08T19:44:24  <besh> Food is spawned symmetrically but i think the tester takes care of that for us if I am not mistaken ..
2011-12-08T19:45:20  <besh> wait that statement doesn't make sense if the engine recieves food in its _visible_ area
2011-12-08T19:46:38  <besh> so i guess , foraging could also benefit from symmetry..
2011-12-08T19:48:44  <LouisMartin> McLeopold: Amstan said sooner (~10h20) that he thought the workers were lowered because of the price. Maybe he didn't know about the tech problem.
2011-12-08T19:51:00  <LouisMartin> McLeopold: the viewer layout is quite attractive. where can I find doc on how to make use of it?
2011-12-08T19:54:17  *** Cyndre <Cyndre!~Cyndre@DOCSIS-182-133.cable-lynx.net> has quit IRC (Quit: Leaving)
2011-12-08T19:55:26  *** bugnuts <bugnuts!~bugnuts@pool-74-108-37-233.nycmny.fios.verizon.net> has left #aichallenge ("Leaving")
2011-12-08T19:55:39  <LouisMartin> Also, does the Java visualizer supports this debug layout?
2011-12-08T19:58:11  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has joined #aichallenge
2011-12-08T20:05:52  *** Anilm3 <Anilm3!~anil@4.Red-79-146-241.dynamicIP.rima-tde.net> has quit IRC (Quit: Lost terminal)
2011-12-08T20:08:04  *** Anilm3 <Anilm3!~anil@4.Red-79-146-241.dynamicIP.rima-tde.net> has joined #aichallenge
2011-12-08T20:11:55  <BenJackson> I'm reworking my strategy layer
2011-12-08T20:12:06  <BenJackson> so I took out almost all of my bot's motivations
2011-12-08T20:12:22  <BenJackson> it doesn't even go toward hills for the most part
2011-12-08T20:12:37  <BenJackson> it might step on one preferentially if it happens to be next to it
2011-12-08T20:12:50  <BenJackson> then I added a slight interest in enemy ants
2011-12-08T20:12:55  <BenJackson> now it's an ant extermination machine
2011-12-08T20:13:16  <BenJackson> finished this 4 player game with total vision of the map and 2 surviving enemies
2011-12-08T20:15:39  <besh> so it is more passive now ?
2011-12-08T20:15:44  *** Migi32 <Migi32!~migi@178-117-102-152.access.telenet.be> has quit IRC (Remote host closed the connection)
2011-12-08T20:16:48  <besh> well I need to add some motivation to my bot. Look at the end of the game here http://ants.fluxid.pl/replay.19863
2011-12-08T20:16:56  <BenJackson> not really
2011-12-08T20:16:58  <BenJackson> just less focused
2011-12-08T20:17:44  <besh> I did the changes you made sometime back and it clearly improved strenght
2011-12-08T20:18:12  <BenJackson> the changes I made?
2011-12-08T20:18:18  <besh> viz a viz. Less attack hill , more exploration, and more enemy ant attacks
2011-12-08T20:18:38  <besh> Hill attack was taking too many ants
2011-12-08T20:18:51  <BenJackson> yeah
2011-12-08T20:19:10  <BenJackson> my current version will do some interesting collapsing on an enemy hill when it's discovered
2011-12-08T20:19:15  <BenJackson> which is great when it's the last
2011-12-08T20:19:17  <BenJackson> and sucks when it's not
2011-12-08T20:19:46  <besh> do you prioritize based on distance to targets ?
2011-12-08T20:20:24  <BenJackson> on all kinds of things, but yes
2011-12-08T20:20:29  <besh> For instance, when I find enemy hill with in 16 unit radius I go for it instead of attacking ants.
2011-12-08T20:21:14  <besh> Also I prefer to attack ants close to my hills with more priority
2011-12-08T20:21:20  <BenJackson> one bug that's fixed in V10 that I saw on aichallenge in V9 was where an ant that was NEXT TO an enemy hill ran off to get food
2011-12-08T20:21:23  <BenJackson> oops
2011-12-08T20:22:00  <besh> yes that could still happen for me
2011-12-08T20:22:10  <BenJackson> I basically rely on finding nearby hills first to set the priority
2011-12-08T20:22:17  <BenJackson> that's one reason I didn't use symmetry detection yet
2011-12-08T20:22:22  <BenJackson> I "find" far away hills too early
2011-12-08T20:22:48  <Anilm3> do the top bots use symmetry detection?
2011-12-08T20:23:07  <besh> But if you have reduced priority for hills, it may come handy
2011-12-08T20:24:19  <BenJackson> Anilm3: I haven't tried to figure it out
2011-12-08T20:24:24  <BenJackson> like I said, I don't, and I was as high as 5th
2011-12-08T20:24:26  *** yoden <yoden!~horstman@cpe-76-190-250-126.neo.res.rr.com> has joined #aichallenge
2011-12-08T20:24:50  <Anilm3> good to know
2011-12-08T20:25:01  <Anilm3> I don't even have a clue as to how that would be implemented
2011-12-08T20:25:40  <BenJackson> I had to read a lot of code
2011-12-08T20:25:44  <BenJackson> the map generators, the engine, etc
2011-12-08T20:27:10  <besh> sysmmetry detection can also be used to reduce path finding _theoretically_
2011-12-08T20:32:13  *** keith__ <keith__!43a1fc8b@gateway/web/freenode/ip.67.161.252.139> has joined #aichallenge
2011-12-08T20:32:23  <Anilm3> I'm still trying to make my bot go a little bit more packed to raze hills and to avoid exploring dead ends
2011-12-08T20:33:23  <Anilm3> I don't even know where to start
2011-12-08T20:33:33  <besh> I try to flock my ants when they are outnumbered by enemy.
2011-12-08T20:34:08  <Anilm3> I see some of the top100 bots and they all go packed and don't get stuck
2011-12-08T20:34:18  <Anilm3> I don't see how they do it
2011-12-08T20:34:28  <thestinger> recursively changing moves
2011-12-08T20:34:45  <Anilm3> what?
2011-12-08T20:34:49  <Anilm3> :O
2011-12-08T20:35:00  <besh> Anilms3: the snake like movements are sideeffects of path finding
2011-12-08T20:35:24  <besh> the head of the snake moves leaving behind an empty space..and so on
2011-12-08T20:36:07  <besh> then when you see an enemy you change course to put maximum attack
2011-12-08T20:36:57  <Anilm3> but path finding usually needs a target doesn't it?
2011-12-08T20:37:24  <besh> yes.
2011-12-08T20:37:53  <Anilm3> In my case
2011-12-08T20:37:57  <Anilm3> when there is no target
2011-12-08T20:38:00  <Anilm3> ants roam freely
2011-12-08T20:38:34  <Anilm3> searching for food and ignoring enemies
2011-12-08T20:38:38  <besh> yes. But when you see a hill all move to one target
2011-12-08T20:38:48  <besh> And they can even flock without any additional code
2011-12-08T20:38:53  <Anilm3> but how can I do that?
2011-12-08T20:39:06  <besh> All ants converge to the same location
2011-12-08T20:39:18  *** keith__ <keith__!43a1fc8b@gateway/web/freenode/ip.67.161.252.139> has quit IRC (Ping timeout: 265 seconds)
2011-12-08T20:39:43  <Anilm3> so if I see a hill I have to move a large number of ants towards it?
2011-12-08T20:39:50  <besh> For exploration (different targets) the ants disperse which is again the desired behaviour
2011-12-08T20:40:05  <besh> yes
2011-12-08T20:40:21  <Anilm3> but pathfinding from every ant to that same hill can be costly
2011-12-08T20:40:34  <BenJackson> jeez there's a lot of spam in the forums
2011-12-08T20:40:41  <BenJackson> I need a macro to report with fewer clicks
2011-12-08T20:40:43  <besh> As a beginner,  infact I moved all my ants
2011-12-08T20:40:59  <Anilm3> I tried that but my bot timed out
2011-12-08T20:41:15  <besh> have you tried bfs ?
2011-12-08T20:41:25  <Anilm3> I did A* from every ant to the target
2011-12-08T20:41:30  <besh> thestinger uses that if I am not mistaken
2011-12-08T20:41:51  <Anilm3> A*?
2011-12-08T20:42:27  <besh> It is more like diffusion without coordination. You generate a bfs table from the hill outwards and then, you ants move towards the hill by minizing distance
2011-12-08T20:43:22  <Anilm3> in order to move a subset of my ants
2011-12-08T20:43:40  <Anilm3> I see
2011-12-08T20:44:05  <besh> what language are you using ?
2011-12-08T20:44:12  <thestinger> besh: I add +3 instead of +1 when going through a tile with one of my ants though
2011-12-08T20:44:13  <Anilm3> c++
2011-12-08T20:44:20  <ikaros> you could sort them by distance so they wont be blocked or implement a "push" system
2011-12-08T20:44:44  <besh> Aha you do that to the bfs table too ?
2011-12-08T20:44:50  <besh> thestinger: ^
2011-12-08T20:45:18  <thestinger> yeah, makes ants spread out or take alternate routes (if they aren't much longer)
2011-12-08T20:45:25  <Anilm3> ikaros: by distance to the hill you mean?
2011-12-08T20:45:32  <ikaros> for example
2011-12-08T20:45:44  <ikaros> or to whatever target you are heading
2011-12-08T20:46:41  <ikaros> though all my bfs based exploration sucked but i guess thats just my fault
2011-12-08T20:47:47  <ikaros> either my ants jittered or the clumped together.. no good distribution :)
2011-12-08T20:48:23  <Anilm3> my exploration actually tends to separate ants too much
2011-12-08T20:48:27  <Anilm3> except on mazes
2011-12-08T20:48:29  <besh> ikaros: the ants do jitter but become straight when they see enemy
2011-12-08T20:48:34  <Anilm3> on mazes they get stuck
2011-12-08T20:48:54  <besh> The first ant to reach the target stops at attack_radius + 2
2011-12-08T20:49:39  <besh> i wouldn't worry about how they move before reaching the target really
2011-12-08T20:49:55  <ikaros> besh :)
2011-12-08T20:50:42  <ikaros> how do you guys decide which tile should be explored?
2011-12-08T20:50:59  <besh> To be honest, I was worried to and tried to make vertical/horizontal lines when i started :)
2011-12-08T20:51:01  <ikaros> i use an times_not_seen attribute but its not so good
2011-12-08T20:51:09  *** QuirionPT <QuirionPT!~quirion@gyf5.kyla.fi> has quit IRC (Quit: Leaving)
2011-12-08T20:51:32  <Anilm3> I try to go to the tiles that haven't been visited in a while
2011-12-08T20:51:44  *** Cyndre <Cyndre!~Cyndre@DOCSIS-182-133.cable-lynx.net> has joined #aichallenge
2011-12-08T20:51:46  <besh> i have that array (visit_count) too but still don't use it
2011-12-08T20:51:50  <ikaros> increasing a counter for each tile when i dont see it
2011-12-08T20:52:21  <Anilm3> something like that
2011-12-08T20:52:24  <ikaros> Anilm3, when i do that and set the period too short the run in circles
2011-12-08T20:52:34  <besh> ikaros: it is better to incrment a counter when you see the tile.
2011-12-08T20:52:44  <ikaros> hm
2011-12-08T20:52:45  <Anilm3> besh: that's how I do it yes
2011-12-08T20:52:52  <ikaros> i had that before
2011-12-08T20:52:54  <ikaros> why is it?
2011-12-08T20:53:06  <Anilm3> I don't increment viewrange
2011-12-08T20:53:13  <Anilm3> just a smaller range
2011-12-08T20:53:23  <Anilm3> arbitrarily picked (worked well)
2011-12-08T20:53:28  <besh> well when I tested the idea, I wanted to move my ant to the least visited tile.
2011-12-08T20:53:31  <ikaros> hm gotta try that
2011-12-08T20:53:56  <Anilm3> ikaros: my ants walk very well on random walks and on some cells
2011-12-08T20:54:03  <Anilm3> but on mazes they get stuck
2011-12-08T20:54:04  <ikaros> besh which is not visible?
2011-12-08T20:54:07  <ikaros> oh hehe
2011-12-08T20:54:39  <Anilm3> I have tried taking into account not only the tiles I haven't visited but the number of ants near that tile
2011-12-08T20:54:44  <Anilm3> but it gets worse somehow
2011-12-08T20:55:22  <ikaros> for mazes this hill climbing thing from the forums was great
2011-12-08T20:55:26  <besh> anilm3: number of enemy ants ?
2011-12-08T20:55:39  <Anilm3> no, number of my own ants
2011-12-08T20:55:39  *** Fandekasp <Fandekasp!~Fandekasp@nt10-ppp1812.west.sannet.ne.jp> has joined #aichallenge
2011-12-08T20:55:43  <ikaros> my old version on the official servers still uses it
2011-12-08T20:55:59  <ikaros> but on rando walks it didnt distribute well enough
2011-12-08T20:56:13  <Anilm3> in ikaros it looks great how it avoids dead ends
2011-12-08T20:56:21  <Anilm3> -in
2011-12-08T20:56:26  <ikaros> yea
2011-12-08T20:56:43  <Anilm3> today I tried some type of "learning"
2011-12-08T20:56:49  <ikaros> but the version is very old and still has the bug: if it discovered the whole map they go all back to base lol
2011-12-08T20:56:53  <Anilm3> but sucked
2011-12-08T20:57:32  <besh> ikaros: i refill unseen tiles when the whole map is visited.
2011-12-08T20:57:44  <ikaros> ok i tried that too
2011-12-08T20:57:49  <besh> I want to replace that with the visit_count array later when I have time
2011-12-08T20:58:04  <ikaros> but often ants went in lines into one direction..
2011-12-08T20:58:13  <ikaros> from the hive
2011-12-08T20:58:25  <besh> unfortunately they do
2011-12-08T20:58:26  <ikaros> and didnt distribute enough.. no idea what i do wrong there
2011-12-08T20:59:07  <besh> well i thought about adding some randomness to the closest tiles, instead of taking the single closest tile as target.
2011-12-08T21:00:00  <ikaros> i used a "ring" of targets as queue for my bfs
2011-12-08T21:00:05  <ikaros> but well
2011-12-08T21:00:07  <ikaros> its 3am
2011-12-08T21:00:11  <ikaros> gotta sleep :)
2011-12-08T21:00:16  <ikaros> good night
2011-12-08T21:00:21  <besh> gn
2011-12-08T21:00:23  *** ikaros <ikaros!~ikaros@dslb-084-059-073-057.pools.arcor-ip.net> has quit IRC (Quit: Ex-Chat)
2011-12-08T21:00:59  <Anilm3> me too
2011-12-08T21:01:05  <besh> gn
2011-12-08T21:01:06  <Anilm3> bye
2011-12-08T21:01:12  *** Anilm3 <Anilm3!~anil@4.Red-79-146-241.dynamicIP.rima-tde.net> has quit IRC (Quit: Lost terminal)
2011-12-08T21:01:17  *** besh <besh!835e7324@gateway/web/freenode/ip.131.94.115.36> has quit IRC (Quit: Page closed)
2011-12-08T21:06:46  *** TheLinker <TheLinker!~TheLinker@host16.190-231-108.telecom.net.ar> has joined #aichallenge
2011-12-08T21:11:48  *** Mystret <Mystret!~Mystret@188.142.37.201> has quit IRC (Ping timeout: 252 seconds)
2011-12-08T21:17:02  *** Paradoxial <Paradoxial!~Paradoxia@pool-108-28-22-94.washdc.fios.verizon.net> has quit IRC (Quit: :)
2011-12-08T21:21:16  *** replore_ <replore_!~replore@203.152.213.161.static.zoot.jp> has joined #aichallenge
2011-12-08T21:44:59  *** besh <besh!835e7324@gateway/web/freenode/ip.131.94.115.36> has joined #aichallenge
2011-12-08T21:58:31  *** Paradoxial <Paradoxial!~Paradoxia@pool-108-28-22-94.washdc.fios.verizon.net> has joined #aichallenge
2011-12-08T22:06:22  *** Paradoxial <Paradoxial!~Paradoxia@pool-108-28-22-94.washdc.fios.verizon.net> has quit IRC (Quit: :)
2011-12-08T22:13:07  *** flowenol <flowenol!LW@bl10-78-112.dsl.telepac.pt> has quit IRC (Ping timeout: 240 seconds)
2011-12-08T22:15:19  *** dmj111 <dmj111!~user@c-98-235-94-211.hsd1.pa.comcast.net> has joined #aichallenge
2011-12-08T22:19:05  *** jeffymj <jeffymj!~scriptz@117.253.185.153> has joined #aichallenge
2011-12-08T22:20:44  *** yoden <yoden!~horstman@cpe-76-190-250-126.neo.res.rr.com> has quit IRC (Quit: Leaving.)
2011-12-08T22:22:13  *** bugnuts <bugnuts!~bugnuts@pool-74-108-37-233.nycmny.fios.verizon.net> has joined #aichallenge
2011-12-08T22:25:01  <jeffymj> quit
2011-12-08T22:25:09  *** jeffymj <jeffymj!~scriptz@117.253.185.153> has left #aichallenge ("Ex-Chat")
2011-12-08T22:27:00  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has quit IRC (Ping timeout: 244 seconds)
2011-12-08T22:27:25  *** AntDroid_ <AntDroid_!781c39b6@gateway/web/freenode/ip.120.28.57.182> has joined #aichallenge
2011-12-08T22:32:16  <LouisMartin> you know you could use a reboot when the ants visualizer has the Deus Ex: Human Revolution Icon in your taskbar
2011-12-08T22:33:24  <besh> R R R _
2011-12-08T22:33:30  <besh> - - - -
2011-12-08T22:33:42  <besh> - - - -
2011-12-08T22:33:59  <besh> - O - O
2011-12-08T22:34:16  <besh> R = your ant  O = opponent ant
2011-12-08T22:35:00  <besh> The correct attack would be to move the top left two ants south, but keep the right one static
2011-12-08T22:35:16  <besh> Assuming static opponent ants. How to achieve this ?
2011-12-08T22:37:19  <LouisMartin> pay them?  ants love money :)
2011-12-08T22:38:34  <LouisMartin> I'm just beginning to develop my own battle system, so I'm afraid I can't afford way more than humor here
2011-12-08T22:38:36  <besh> they love sugar :) My combat moves all three ants down and I think it can help to improve this
2011-12-08T22:39:49  <besh> I wonder if the influence methods discussed would help here
2011-12-08T22:39:57  <LouisMartin> well, what about less sugar where your ant is sure to get killed without gain?
2011-12-08T22:41:37  <besh> The third ant will find no sugar by moving down, but two angry enemy attacks
2011-12-08T22:42:04  <besh> You are right, i should start thinking interms of sugar :)
2011-12-08T22:42:13  *** sigh <sigh!~sigh@ppp114-153.static.internode.on.net> has quit IRC (Remote host closed the connection)
2011-12-08T22:42:55  <thestinger> my latest attempt at combat involves lying to my ants about how many enemies there are :P
2011-12-08T22:43:32  <besh> do you increase the sugar level so they would be motivated to attack ?
2011-12-08T22:44:22  *** LouisMartin_ <LouisMartin_!6017820a@gateway/web/freenode/ip.96.23.130.10> has joined #aichallenge
2011-12-08T22:44:29  *** LouisMartin <LouisMartin!6017820a@gateway/web/freenode/ip.96.23.130.10> has quit IRC (Ping timeout: 265 seconds)
2011-12-08T22:44:33  <LouisMartin_> 1, 2, 1, 2 testing
2011-12-08T22:44:41  <besh> thestinger: seriously I have trouble solving the above example
2011-12-08T22:44:54  *** cyphase <cyphase!~cyphase@unaffiliated/cyphase> has quit IRC (Ping timeout: 248 seconds)
2011-12-08T22:45:19  <thestinger> my old method was just to attack with everything
2011-12-08T22:45:22  <thestinger> then try cancelling orders
2011-12-08T22:45:31  <thestinger> which would result in the left two going down and the right one not going down
2011-12-08T22:45:36  <LouisMartin_> besh: sorry if I repeated this many times, but my computer is bugging tight and I don't know if I've resent this many times over.. strange bugs.. anyways ->
2011-12-08T22:45:38  <LouisMartin_> the implementation of this depends on how your system works, but a quick check such as "if I don't sacrifice this ant, is the outcome (gain) different?"  before moving wouldn't be costly I'd think
2011-12-08T22:46:39  <thestinger> going to bed now, cya
2011-12-08T22:47:09  <besh> what if there was another ant north of the 3rd ant. I mean redoing a move woud triger another undo
2011-12-08T22:47:19  <besh> undoing*
2011-12-08T22:47:29  <besh> ok cya
2011-12-08T22:47:36  *** amstan <amstan!~alex@69-196-130-230.dsl.teksavvy.com> has joined #aichallenge
2011-12-08T22:47:36  *** ChanServ sets mode: +o amstan
2011-12-08T22:49:13  *** antimatroid <antimatroid!~nick@141.114-27-211.dynamic.dsl.mel.iprimus.net.au> has quit IRC (Ping timeout: 244 seconds)
2011-12-08T22:50:05  *** LouisMartin_ <LouisMartin_!6017820a@gateway/web/freenode/ip.96.23.130.10> has quit IRC (Quit: Page closed)
2011-12-08T22:55:28  *** Antimony <Antimony!~chatzilla@lawn-143-215-61-158.lawn.gatech.edu> has joined #aichallenge
2011-12-08T22:58:55  <dmj111>         result = []
2011-12-08T22:59:33  *** Fandekasp <Fandekasp!~Fandekasp@nt10-ppp1812.west.sannet.ne.jp> has quit IRC (Ping timeout: 244 seconds)
2011-12-08T23:02:45  *** sigh <sigh!~sigh@ppp114-153.static.internode.on.net> has joined #aichallenge
2011-12-08T23:03:23  *** bmh <bmh!ada4e181@fsf/member/bmh> has joined #aichallenge
2011-12-08T23:03:48  *** dvladim <dvladim!~dvladim@85.113.131.65> has joined #aichallenge
2011-12-08T23:05:47  <bmh> hmm... make my BFS code more generic, or start on fighting
2011-12-08T23:07:56  *** delt0r_ <delt0r_!~bob@91-113-94-210.adsl.highway.telekom.at> has quit IRC (Ping timeout: 244 seconds)
2011-12-08T23:11:33  *** besh <besh!835e7324@gateway/web/freenode/ip.131.94.115.36> has quit IRC (Quit: Page closed)
2011-12-08T23:12:26  *** cyphase <cyphase!~cyphase@unaffiliated/cyphase> has joined #aichallenge
2011-12-08T23:12:41  <Antimony> Ug, I keep finding bugs in my symmetry code
2011-12-08T23:12:54  <Antimony> I've found at least 3 bugs in release versions
2011-12-08T23:13:10  <bmh> Antimony: you're doing symmetry? neat!
2011-12-08T23:13:45  <bmh> what 's your technique?
2011-12-08T23:14:50  <Antimony> what do you mean?
2011-12-08T23:15:58  <bmh> How are you solving it? How do you deal with rotation/mirroring?
2011-12-08T23:16:54  <Antimony> well for one thing, I assume rotation can only happen on square maps
2011-12-08T23:17:28  <bmh> that's a nice simplifying assumption
2011-12-08T23:18:12  <Antimony> anyway, I hope that it will give me a edge, but I haven't really noticed any benefits yet
2011-12-08T23:18:23  <Antimony> it wasn't as effective as I'd hoped
2011-12-08T23:18:56  <bmh> currently my ants blindly rush other ants
2011-12-08T23:19:01  <bmh> it's... suboptimal
2011-12-08T23:20:18  *** delt0r_ <delt0r_!~bob@178-190-200-11.adsl.highway.telekom.at> has joined #aichallenge
2011-12-08T23:20:37  <Antimony> why would you rush other ants?
2011-12-08T23:21:15  <bmh> Antimony: because I have no fight code
2011-12-08T23:21:50  <Antimony> Anyway, the basic idea is to find all possible syemmtries and then each time you see a new square, remove the ones that don't fit. But unfortunately the method is a little slow
2011-12-08T23:22:48  <Antimony> Some times, my bot doesn't even get done processing the starting data before it finds the first hill
2011-12-08T23:22:55  <bmh> slow, really? :)
2011-12-08T23:22:58  <a1k0n> wow, the contest site is cruising now
2011-12-08T23:27:03  <bmh> Random Number Generator says: Genericize your search code.
2011-12-08T23:27:33  *** dvladim <dvladim!~dvladim@85.113.131.65> has quit IRC (Ping timeout: 248 seconds)
2011-12-08T23:27:38  *** br3akd0wn <br3akd0wn!~mint@c-98-193-68-113.hsd1.il.comcast.net> has joined #aichallenge
2011-12-08T23:28:01  <amstan> Antimony: hash
2011-12-08T23:28:21  <Antimony> ?
2011-12-08T23:28:30  <amstan> for symmetry detection
2011-12-08T23:28:35  *** br3akd0wn <br3akd0wn!~mint@c-98-193-68-113.hsd1.il.comcast.net> has quit IRC (Client Quit)
2011-12-08T23:28:54  <Antimony> nah, I think my way is better even if it is slow
2011-12-08T23:29:15  <Antimony> and I do it at the end of turn after moves are calculated so at least it isn't hurting me
2011-12-08T23:29:36  <amstan> also... once you find your first hill it's really easy
2011-12-08T23:29:53  <amstan> the challenge is to do this before you find the hill, so you know where they are before the enemy has seen you
2011-12-08T23:30:07  <BenJackson> I keep "timing out" on fluxid
2011-12-08T23:30:20  <BenJackson> time taken: 185.627ms
2011-12-08T23:30:22  <amstan> BenJackson: turn off those torrents, lol
2011-12-08T23:30:45  <BenJackson> my ping to fluxid.pl is about 200ms
2011-12-08T23:31:19  <BenJackson> some packet loss in mid europe
2011-12-08T23:31:27  <_flag> BenJackson: fluxid times me out all the time too
2011-12-08T23:31:43  <_flag> That server is kind of known for its problems, I think there may just be a ton of people using it
2011-12-08T23:32:36  <BenJackson> this bot has most of the logic removed, too
2011-12-08T23:32:42  <BenJackson> so I was worried if it *really* timed uot!
2011-12-08T23:33:06  <bmh> my server isn't very beefy, but you could run your real bot there?
2011-12-08T23:33:19  <BenJackson> in 20 games the worst turn was 310ms
2011-12-08T23:33:48  <BenJackson> bmh: you want me to run my current contest bot on your server?
2011-12-08T23:34:17  <bmh> BenJackson: go for it. If fluxid is overrun, I have no qualms with 'production' bots running
2011-12-08T23:34:26  <BenJackson> oh, no
2011-12-08T23:34:29  <Antimony> my bot can generally only find hills before seeing them on a 2 player map
2011-12-08T23:34:35  <BenJackson> I just did 20 runs with some modified logic to see what it did
2011-12-08T23:34:43  <Antimony> because you usally have to explore almost half the map before you can be sure
2011-12-08T23:34:49  <BenJackson> I just wanted to see what it did with most of the "rules" removed
2011-12-08T23:34:53  <BenJackson> to see which rules are worth putting back
2011-12-08T23:35:02  <bmh> BenJackson: there is one game with bj_v4 v. skynet. Skynet should have won, but I turned off my laptop and went to bed :)
2011-12-08T23:35:14  <BenJackson> Antimony: you guys are going to kick yourselves at the end ;-)
2011-12-08T23:35:35  *** avdg <avdg!~avdg@78-22-168-200.access.telenet.be> has quit IRC (Quit: Leaving.)
2011-12-08T23:35:45  <Antimony> I doubt your symetry method is better than mine
2011-12-08T23:36:00  <BenJackson> :)
2011-12-08T23:36:02  <Antimony> well apart from optimization
2011-12-08T23:36:12  <Antimony> mine's good, it's just too slow
2011-12-08T23:36:14  <BenJackson> :)
2011-12-08T23:37:07  *** amstan <amstan!~alex@69-196-130-230.dsl.teksavvy.com> has quit IRC (Ping timeout: 240 seconds)
2011-12-08T23:37:48  *** avdg <avdg!~avdg@78-22-168-200.access.telenet.be> has joined #aichallenge
2011-12-08T23:37:49  *** amstan_ <amstan_!~amstan@69-196-130-230.dsl.teksavvy.com> has quit IRC (Read error: Operation timed out)
2011-12-08T23:37:54  *** TheLinker <TheLinker!~TheLinker@host16.190-231-108.telecom.net.ar> has quit IRC (Quit: Bye)
2011-12-08T23:42:18  *** amstan <amstan!~alex@69-196-130-230.dsl.teksavvy.com> has joined #aichallenge
2011-12-08T23:42:18  *** ChanServ sets mode: +o amstan
2011-12-08T23:46:05  *** Scryer <Scryer!4a645367@gateway/web/freenode/ip.74.100.83.103> has joined #aichallenge