2010-11-07T00:07:24  <Prillicy> yeah, I know
2010-11-07T00:08:00  <Prillicy> interesting, there is no planet.Fleets() function
2010-11-07T00:08:08  <Prillicy> am I ok to create that extension method?
2010-11-07T00:08:14  <McLeopold> Yes
2010-11-07T00:08:24  <McLeopold> If fact, I would suggest re-writing the starter pack.
2010-11-07T00:08:37  <McLeopold> I don't create any fleet objects.
2010-11-07T00:08:37  <Prillicy> eventually
2010-11-07T00:08:41  <Prillicy> but first I have to learn it
2010-11-07T00:08:49  <Prillicy> why not?
2010-11-07T00:08:53  <Prillicy> fleets are vital
2010-11-07T00:08:58  <McLeopold> I just index the fleets by the destination planet.
2010-11-07T00:09:01  <Prillicy> knowing where your enemies are is important
2010-11-07T00:09:03  <Prillicy> ah, I see
2010-11-07T00:09:13  <Prillicy> I was thinking the other way arond
2010-11-07T00:09:15  <Prillicy> source planet
2010-11-07T00:09:18  <McLeopold> I have an array called, incoming_enemy
2010-11-07T00:09:24  <Prillicy> lol
2010-11-07T00:09:26  <McLeopold> I just add them all up
2010-11-07T00:09:47  <Prillicy> yeah, I'm not as concerned about them as I am about me
2010-11-07T00:09:48  <McLeopold> Well, once they leave the source, you should only care about where they are going
2010-11-07T00:09:52  <Prillicy> not yet at least
2010-11-07T00:12:47  <Prillicy> what is this AddShips() method?
2010-11-07T00:14:29  <McLeopold> not sure, I don't use the starter pack
2010-11-07T00:15:34  <contestbot> Top 10 players: bocsimacko(4311), shangas(4167), Raschi(4130), bix0r4ever(4060), george(4044), dmj111(4034), Xrillo(4022), rsergio(4016), Slin-.-(4006), GreenTea(4005)
2010-11-07T00:41:29  *** demerzel_ <demerzel_!488039c1@gateway/web/freenode/ip.72.128.57.193> has joined #aichallenge
2010-11-07T00:41:50  <demerzel_> lol where were you when everybody thought i was crazy for not having a fleet object McLeopold
2010-11-07T00:47:29  <McLeopold> I missed that
2010-11-07T00:47:36  <McLeopold> You're not crazy
2010-11-07T00:47:45  <demerzel_> i posted my old C# framework and it uses that system
2010-11-07T00:47:52  <demerzel_> destination fleet arrays by turn
2010-11-07T00:47:57  <McLeopold> I can't think of a good reason why I need one.
2010-11-07T00:48:10  <demerzel_> nope - if it weren't for the screwy neutral thing you woudln't need them at all
2010-11-07T00:48:15  <McLeopold> were there any good arguments for?
2010-11-07T00:48:31  <demerzel_> no
2010-11-07T00:49:07  <McLeopold> It's a classic case of oop programmers unable to think outside the box.
2010-11-07T00:49:22  <McLeopold> I really detest oop.
2010-11-07T00:49:48  <demerzel_> i don't hate it but if it's not needed it's not needed
2010-11-07T00:49:49  <McLeopold> Everytime I see a getter and setter that only gets and sets one private variable, I want to vomit.
2010-11-07T00:49:55  <demerzel_> haha
2010-11-07T00:50:07  <demerzel_> the only time i really do that is if it's going to be portable and usually at work
2010-11-07T00:50:19  <McLeopold> I did learn to appreciate java, eventually, but the IDE better write that crap for me.
2010-11-07T00:50:36  <demerzel_> what lang do you use
2010-11-07T00:50:45  <demerzel_> i am trying to set up a linked node system for a tree
2010-11-07T00:50:46  <McLeopold> I use the all.
2010-11-07T00:50:47  <demerzel_> in c
2010-11-07T00:51:06  <McLeopold> that might be fun
2010-11-07T00:51:08  <demerzel_> well setting up the struct itself is easy, just trying to frame it in my mind
2010-11-07T00:51:13  <demerzel_> it's pretty damn easy in c actually
2010-11-07T00:51:31  <McLeopold> I haven't done c in over ten years.
2010-11-07T00:51:33  <demerzel_> struct a has a pointer to struct b
2010-11-07T00:51:56  <demerzel_> i might have to forward dec it to do pointers to same struct not sure yet
2010-11-07T00:52:11  <demerzel_> i love C
2010-11-07T00:52:17  <demerzel_> i was doing c++ but just dropped the ++
2010-11-07T00:52:30  <demerzel_> mostly working with pure numbers it doesn't make sense
2010-11-07T00:53:03  <McLeopold> Yeah, I prefer just pure numbers and algorithms.
2010-11-07T00:53:13  <McLeopold> Launguage is secondary.
2010-11-07T00:53:15  <demerzel_> well this is definitely showing my lack of formal education lol
2010-11-07T00:53:25  <demerzel_> i am having difficulty framing the game theory algorithms
2010-11-07T00:53:40  <demerzel_> been forever since i had to read formulas
2010-11-07T00:53:40  <McLeopold> are you doing alpha beta?
2010-11-07T00:53:55  <demerzel_> not sure yet - i have yet to define exactly what a 'move' is
2010-11-07T00:54:21  <demerzel_> also if multiple moves are available on turn x then how to prune doubling effort etc
2010-11-07T00:54:29  <demerzel_> and if i can use hungarian how to set up the values blahblah
2010-11-07T00:54:55  <McLeopold> well, from a simplistic point of view...
2010-11-07T00:55:03  <McLeopold> ignoring the giant move tree...
2010-11-07T00:55:04  <demerzel_> (mine)
2010-11-07T00:55:32  <McLeopold> A move would be all combinations of fleet movements that each player can make on a given turn.
2010-11-07T00:55:42  <demerzel_> oh i abandoned that idea haha
2010-11-07T00:56:03  <demerzel_> at the very least i'll limit it to moves that have some kind of gain
2010-11-07T00:56:04  <McLeopold> For a 2 planet map (lame) with 100 ships, there would be 101 moves for each player...
2010-11-07T00:56:21  <McLeopold> so 101*101 total "moves" or new games states.
2010-11-07T00:56:25  <demerzel_> yea i used to run it like that
2010-11-07T00:56:43  <demerzel_> my first C# bot ran 1-n ships available to all planets(including mine)
2010-11-07T00:56:45  <McLeopold> If we set up some cloud computing, we could do that
2010-11-07T00:56:56  <demerzel_> only this turn obv lol
2010-11-07T00:57:23  <demerzel_> i doubt that non-gain moves have any advantage - beyond staging which can be looked for
2010-11-07T00:57:32  <demerzel_> i am not going to tackle that initially tho
2010-11-07T00:57:49  <McLeopold> I think you could limit it by considering combinations of targets.
2010-11-07T00:57:59  <demerzel_> that's my tentative idea
2010-11-07T00:58:08  <demerzel_> i am working out how to standardize the value system
2010-11-07T00:58:26  <demerzel_> considering that i could make move x which has a return in 8 turns vs move y which has a return in 10 turns
2010-11-07T00:58:35  <McLeopold> So if there are 4 planets on the map, there 16 combinations of planets to try to go fore.
2010-11-07T00:59:05  <demerzel_> well that can be limited down more than that
2010-11-07T00:59:21  <demerzel_> if enemy presence > my presence, abandon
2010-11-07T00:59:24  *** Utkarsh <Utkarsh!~admin@117.201.80.20> has quit IRC (Ping timeout: 265 seconds)
2010-11-07T00:59:38  <McLeopold> maybe.  if you can't afford them all, the don't try the combination of all 4
2010-11-07T01:00:02  <demerzel_> well but on a map with 4 planets then the combinations go from 4! to 2!
2010-11-07T01:00:29  <demerzel_> and in the case of the 4! it might go down farther than that because taking 1 precludes taking 2 others etc
2010-11-07T01:00:33  <McLeopold> I think that is too limiting
2010-11-07T01:00:59  <demerzel_> in what situation do you see value in attackign a target that can be countered with 100% probability?
2010-11-07T01:01:19  <McLeopold> waiting to snipe
2010-11-07T01:01:33  <demerzel_> ?
2010-11-07T01:02:09  <McLeopold> but that may be solved if you get the alpha beta working
2010-11-07T01:02:36  <McLeopold> actually, you're right, only go for planets with less enemy presence.
2010-11-07T01:02:49  *** Frontier <Frontier!~Marco_Lei@dslb-088-071-242-074.pools.arcor-ip.net> has joined #aichallenge
2010-11-07T01:03:02  <McLeopold> the next turn, if the enemy commits too much, you should see their influence drop and then you can consider the planets
2010-11-07T01:03:55  <demerzel_> it might be feasible to look at >= not sure yet
2010-11-07T01:04:27  <demerzel_> this bot only goes for enemy presence < my presence plus sends along possible enemy presence
2010-11-07T01:04:27  <demerzel_> http://spz.nofate.com/planetwars/visualizer/?game_id=888912&server=tcp
2010-11-07T01:04:30  <demerzel_> doesn't do too bad
2010-11-07T01:04:31  <McLeopold> unless the num of ships on a neural is less than or equal to the growth, always wait
2010-11-07T01:04:58  <McLeopold> map5
2010-11-07T01:05:06  <demerzel_> the other thing i'm considering is followthrough
2010-11-07T01:05:23  <demerzel_> ie if movex->movey made sense last turn should i tell it to consider movey first or make it recalc the whole thing
2010-11-07T01:05:29  *** Utkarsh <Utkarsh!~admin@117.201.90.56> has joined #aichallenge
2010-11-07T01:05:39  <demerzel_> and if so do i lose anything
2010-11-07T01:05:42  <Frontier> dmerzel_: You took a lot of beating for your source code release. What I wondered about is that a simple one-line-logic makes it up to rank 4xx. The other bots can't all have a worse logic, right?
2010-11-07T01:05:44  <McLeopold> that sound like it works for chess, but not this
2010-11-07T01:06:37  <demerzel_> Frontier: it's pure one-target cost effectiveness ratio and yes they can
2010-11-07T01:06:58  <demerzel_> plus the advanced state stuff is well beyond most people's comprehension and/or willing effort
2010-11-07T01:07:29  <Frontier> so... 90% of the participants are ... I don't know ...
2010-11-07T01:07:43  <demerzel_> consider the starter package - it cant' tell you jack about anything beyond this turn plus gives a model of source % which people modify to start with
2010-11-07T01:07:58  <McLeopold> If my original bot is still in the top 100, then there are not that many good bots out there.
2010-11-07T01:07:58  <demerzel_> a bad bad start
2010-11-07T01:09:16  <demerzel_> Frontier: you are assuming that everyone that enters is a real developer, there are 'real developers' who can't develop
2010-11-07T01:09:19  <Frontier> i guess my original would be at 300 by now. so you put in some foresight into that bot demerzel_? Some framework that is better suited than the starter package and that is what actually makes that bot good?
2010-11-07T01:09:53  <demerzel_> plus the efficiency required precludes a lot of crutches people are used to like lambdas etc - at least not to any depth
2010-11-07T01:10:24  <demerzel_> Frontier: yes - my state calc is really all the advantage i ever had, my logic never went beyond cost-effectiveness until recently
2010-11-07T01:10:25  <Frontier> some guy i know from a forum is @rank 500 right now and he thought he achieved something, haha
2010-11-07T01:11:19  <demerzel_> well it means he's not an idiot - and if he's still using the starter pack then even more so cuz that thing is not well-suited to the kind of calcs you really need without a ton of mod
2010-11-07T01:12:18  <demerzel_> thats kinda why i released that even if nobody will ever use it - i came to my design after several false starts based on bad assumptions, and you can't really avoid that when info is scarce and the starter packs barely work
2010-11-07T01:12:30  <demerzel_> i'm not an information hoarder haha
2010-11-07T01:13:37  <demerzel_> now when i get into my game trees and shit you're on your own lol
2010-11-07T01:15:35  <contestbot> Top 10 players: bocsimacko(4315), shangas(4173), Raschi(4132), bix0r4ever(4052), george(4046), GreenTea(4045), dmj111(4036), Xrillo(4024), rsergio(4022), Slin-.-(4007)
2010-11-07T01:16:14  <demerzel_> i also remember early testing was a nightmare because i barely knew how to test console apps without some kind of wrap and i built a game engine and a proxy bot etc till i just came up with 'why not plug in tcp directly'
2010-11-07T01:16:33  <demerzel_> then just pipe stdin to a file, if it fucks up then flip a switch and read from file
2010-11-07T01:18:35  <Frontier> i see, and i didn't have an opinion on sharing framework code until now. but i think the advantage that c# beginners can have now above users of other languages is unfair. it is a big time saver at least, when it puts you on rank 400 without much effort.
2010-11-07T01:20:20  <demerzel_> assuming they even go to the forums, d/l my code and put the time in to understand it
2010-11-07T01:20:56  <demerzel_> rank 400 means nothing
2010-11-07T01:21:17  <Frontier> if noone downloads and uses it, then why did you put it up? ^^
2010-11-07T01:21:22  <McLeopold> Frontier: maybe I'll release my python framework so it's more fair. :P
2010-11-07T01:21:30  <demerzel_> because i'm nice
2010-11-07T01:21:42  <demerzel_> it would just go to waste if i didn't
2010-11-07T01:21:51  <demerzel_> i'm not even using it lol
2010-11-07T01:22:23  <Frontier> do you know what a script kiddie is?
2010-11-07T01:22:52  <amstan> imho frameworks and other code is not really an issue
2010-11-07T01:22:52  <McLeopold> script kiddies are people too
2010-11-07T01:23:28  <amstan> if one's smart he can just skip those steps and advance more
2010-11-07T01:23:48  <demerzel_> lol script kiddie
2010-11-07T01:23:51  <amstan> if you want to avoid spoilers do something else.. like watch a movie
2010-11-07T01:23:56  <Frontier> script kiddies are like gollum, they search the forums for their treasure - ready to use code for high ranking bots - and ruin other peoples work on getting above rank 500
2010-11-07T01:24:21  <demerzel_> i get that people put tons of time into their stuff and don't like that other people might get something comparable for free but that's really a close-minded way to look at it
2010-11-07T01:24:41  <demerzel_> this isn't even a free ride it probably has bugs i didn't get to test it a lot
2010-11-07T01:24:43  <amstan> Frontier: ruin other people's work? lol.. they had exactly the same chance of looking though the forums findings good ideas
2010-11-07T01:25:11  <demerzel_> or irc haha
2010-11-07T01:25:22  <demerzel_> that's where i get my script kiddie ideas
2010-11-07T01:25:24  <Frontier> it is a matter of dope or not to dope, see how the bicycle sport is going down
2010-11-07T01:25:59  <McLeopold> Frontier: I agree with you, that it is unfair.  But this contest is not to see who can produce the best bot on their own effort.  It is simple to see who can produce the best bot with any type of effort.
2010-11-07T01:26:00  <amstan> it's a matter of treating those beggining steps like a spoiler... but this is a contest, not a book
2010-11-07T01:26:25  <demerzel_> lol so when somebody posted the hungarian python implementation that was blowing the competition too
2010-11-07T01:26:26  <McLeopold> Frontier: It is going to be tough to keep up because of released code.
2010-11-07T01:26:45  <demerzel_> no it won't have you seen the latest top 10 worry about them
2010-11-07T01:26:51  <amstan> it's either experiencing the contest slowly, while not getting any outside help, or gathering as much help as possible and using it to your advantage
2010-11-07T01:27:40  <demerzel_> somebody starting today has zero chance without outside help
2010-11-07T01:27:43  <amstan> i won the rps challenge by writing like 50 lines of code that was made by me only
2010-11-07T01:27:51  <amstan> the rest was greenberg..
2010-11-07T01:27:52  <amstan> lol
2010-11-07T01:27:55  *** plediii <plediii!~plediii@adsl-99-185-124-47.dsl.hstntx.sbcglobal.net> has quit IRC (Ping timeout: 240 seconds)
2010-11-07T01:27:55  <McLeopold> It's closer to the real world as well.  You don't reinvent the wheel when writing commercial software.  You apapt proven solutions.
2010-11-07T01:28:03  <amstan> exactly
2010-11-07T01:28:14  <McLeopold> amstan: but I still don't like it. :P
2010-11-07T01:28:22  <Frontier> hey, freeware can kill commercial companies
2010-11-07T01:28:26  <amstan> life's not fair
2010-11-07T01:28:33  <demerzel_> McLeopold: i reinvent the wheel anyway
2010-11-07T01:28:50  <demerzel_> but i realize it's dumb
2010-11-07T01:28:51  <McLeopold> demerzel_: So do I.  My wheels are always better.
2010-11-07T01:28:58  <amstan> McLeopold: +1
2010-11-07T01:29:15  <amstan> McLeopold: are yours spherical?
2010-11-07T01:29:22  <demerzel_> but i figure the next time i see a wheel i know exactly whether it's good or not ;)
2010-11-07T01:30:12  <demerzel_> it's really gonna suck when this is over if tcp goes away etc
2010-11-07T01:30:13  <McLeopold> Only if it's in the requrements.  I prefer reuleaux triangles.
2010-11-07T01:30:22  <demerzel_> this has been one of my more time-intensive effort
2010-11-07T01:30:51  <McLeopold> I will keep tcp alive for ever if I have to.
2010-11-07T01:31:01  <McLeopold> This stuff is way too fun.
2010-11-07T01:31:19  <Frontier> demerzel_: I guess there wont be much traffic on tcp anymore, you would have to stay in contact with others. and while you are at it.. set up your own tcp server
2010-11-07T01:31:50  <Frontier> It could be a peer-to-peer game.
2010-11-07T01:32:57  <demerzel_> hell i've thought about setting up my own tcp server
2010-11-07T01:33:05  <demerzel_> i have a vps running minecraft which i haven't played in like 2m onths
2010-11-07T01:33:37  <Frontier> What about a master game server, where you can join running games of up to x players and then start a tournament on selected maps?
2010-11-07T01:33:40  <demerzel_> not a bad idea maybe keep it running like that tank war game
2010-11-07T01:33:47  <demerzel_> not like it costs that much to keep text strings in/out
2010-11-07T01:33:50  <Prillicy> I'm gonna talk, but I'm mostly just talking to think out lout if that's alright
2010-11-07T01:34:35  <Prillicy> I need to build an array of planets that is sorted in the order of the distance they are from each of my planets...so my planets variable needs an extension with this array and the array's key needs to be distances...
2010-11-07T01:34:52  <Prillicy> hrm....
2010-11-07T01:34:56  <Frontier> this master game server could also list tron, rps, go, whatever programming game and a client would connect to and visualize the running game
2010-11-07T01:35:08  <Prillicy> I think I am going to have to rewrite this starter pack.
2010-11-07T01:35:23  <Frontier> Prillicy: there is a starter package in c# for that! -.-
2010-11-07T01:35:35  <Prillicy> I have the starter package...
2010-11-07T01:36:08  <Frontier> but there is one that even calculates future planet states for you
2010-11-07T01:36:15  <demerzel_> C# has a builtin array.sort(keys, items) arrays
2010-11-07T01:36:26  <demerzel_> i'm sure java has something similar
2010-11-07T01:36:30  <demerzel_> made distance sorting a breeze
2010-11-07T01:36:57  <Prillicy> brilliant
2010-11-07T01:37:01  <Frontier> Java has SortedList
2010-11-07T01:38:00  <demerzel_> sorted list isn't arrays
2010-11-07T01:38:12  <demerzel_> c# has that too but i avoid collections unless i actually need them
2010-11-07T01:39:58  <Frontier> you are right, the access isn't as fast as in an array. i prefer arrays over lists, too
2010-11-07T01:44:37  <demerzel_> i am like the palpatine of programming languages
2010-11-07T01:44:47  <demerzel_> come to the dark side, feel the power of C
2010-11-07T01:46:34  <demerzel_> i just squealed like a little girl cuz i made a 2-level tree and it just spit out a crapton of evals
2010-11-07T01:46:42  <demerzel_> it's ALIVE
2010-11-07T01:50:56  *** hellman <hellman!~hellman@nat232-248-205-109.tvoe.tv> has joined #aichallenge
2010-11-07T01:51:52  <Frontier> demerzel_: lol, all will be good. so DO you use c now? and what is a 2-level-tree?
2010-11-07T01:52:40  <McLeopold> demerzel_: http://visualizer.naktibalda.lt/522
2010-11-07T01:53:00  <demerzel_> Frontier: yes it's iso-c - 2 level = 2-ply i guess
2010-11-07T01:53:15  <Frontier> bocsimacko is probably winning because he is confusing his opponent with strongly varying ships counts on his planets, giving the enemy no clear targets
2010-11-07T01:53:34  <demerzel_> McLeopold: i thought antbot < scotbot
2010-11-07T01:53:41  <McLeopold> nope
2010-11-07T01:53:52  <McLeopold> well, on tcp it has been, because of bugs
2010-11-07T01:53:59  <McLeopold> AntBot is my best
2010-11-07T01:54:20  <Frontier> reminds me of: http://en.wikipedia.org/wiki/Pholcidae#Threat_response
2010-11-07T01:55:00  <Frontier> demerzel_: So you can look 2 turns ahead or was that a joke?
2010-11-07T01:55:56  <demerzel_> Frontier: not 2 turns, 2 moves
2010-11-07T01:56:07  <McLeopold> what's the difference?
2010-11-07T01:56:42  <McLeopold> wait, one fleet order per node?
2010-11-07T01:56:50  <Frontier> corrreeeect
2010-11-07T01:57:34  <Frontier> I would like to do the teal'c 'that is correct' but it doesn't sound as good on irc
2010-11-07T01:57:44  *** Utkarsh <Utkarsh!~admin@117.201.90.56> has quit IRC (Ping timeout: 250 seconds)
2010-11-07T01:57:47  <demerzel_> this will probably be the last time i am open about my tree since it is rapidly approaching actual usability but this was the log: http://pastebin.com/62L3JtKF
2010-11-07T01:58:34  <Frontier> CRAZY: it looks like what i pasted yesterday and put up on irc, just to regret that sigh got an inspriation from it
2010-11-07T01:58:57  <McLeopold> atk?
2010-11-07T01:59:04  <demerzel_> as opposed to DEF ;)
2010-11-07T01:59:17  <McLeopold> ?
2010-11-07T01:59:24  <demerzel_> it's a test map so nobody is taking my planets
2010-11-07T01:59:37  <demerzel_> attack = capture vs def = defend
2010-11-07T01:59:45  <McLeopold> oh
2010-11-07T01:59:47  <McLeopold> I'm slow
2010-11-07T01:00:20  <amstan> omg! yes! another chance to use 1AM
2010-11-07T01:00:25  <Frontier> demerzel_: I started out the same, with an incative enemy and I have to find a way to incorporate the enemy now.
2010-11-07T01:00:38  <demerzel_> anyway i'm done for today i think - state seems rock-solid finally and i'm actually getting somewhere with logic
2010-11-07T01:00:51  <demerzel_> a good day!
2010-11-07T01:01:00  <demerzel_> amstan: 1AM?
2010-11-07T01:01:19  <amstan> demerzel_: daylight savings
2010-11-07T01:01:28  <amstan> demerzel_: see http://contestbot.hypertriangle.com/channellogs/%23aichallenge/%23aichallenge.2010-11-07.log
2010-11-07T01:01:49  <demerzel_> what that's today?  damn
2010-11-07T01:02:50  <demerzel_> wow so it's actually 2
2010-11-07T01:02:51  *** Utkarsh <Utkarsh!~admin@117.201.87.97> has joined #aichallenge
2010-11-07T01:02:53  <demerzel_> haha i had no idea
2010-11-07T01:02:59  <McLeopold> amstan: how's the terrain code?  got any pictures of good results?
2010-11-07T01:03:10  <demerzel_> gn :)
2010-11-07T01:03:13  *** demerzel_ <demerzel_!488039c1@gateway/web/freenode/ip.72.128.57.193> has quit IRC (Quit: Page closed)
2010-11-07T01:03:15  <amstan> McLeopold: didn't do much yet, still trying to fix some performance issues
2010-11-07T01:03:39  <McLeopold> http://72.44.46.68/canvas?game_id=889752
2010-11-07T01:03:51  <McLeopold> My bot looks dumb at first, but just wait...
2010-11-07T01:04:09  <McLeopold> I've never been so proud. (sniff)
2010-11-07T01:05:47  <McLeopold> amstan: you never sent me a picture of why cliff are bad
2010-11-07T01:11:32  <amstan> oh man.. this is the longest 3mb upload i ever did
2010-11-07T01:13:40  <McLeopold> amstan: what are you make with the terrains?  AmstanCraft?
2010-11-07T01:13:52  <McLeopold> MindStan?
2010-11-07T01:13:57  <amstan> sauerbraten maps
2010-11-07T01:14:04  <Prillicy> my bot isn't compiling...is this a common issue?
2010-11-07T01:14:20  <amstan> McLeopold: http://cubicmatter.net/forums/viewtopic.php?f=10&t=54
2010-11-07T01:14:22  <McLeopold> If you mean on the official site with C#, yes.
2010-11-07T01:14:32  <amstan> Prillicy: errors are common issues in programming, yes
2010-11-07T01:14:41  <Prillicy> no, no error
2010-11-07T01:14:42  <amstan> oh, i wasn't aware of that
2010-11-07T01:14:43  <Prillicy> just blank screen
2010-11-07T01:15:35  <contestbot> Top 10 players: bocsimacko(4307), shangas(4156), Raschi(4129), bix0r4ever(4063), rsergio(4034), Xrillo(4019), GreenTea(4019), sin_sun(4016), george(4014), dmj111(4008)
2010-11-07T01:16:25  <Prillicy> http://pastebin.com/3uAM81Ex <---doesn't compile
2010-11-07T01:16:31  <McLeopold> amstan: what am I suppost to look at?
2010-11-07T01:16:50  <amstan> McLeopold: the bottom stuff mostly.. it's before we went on editing corners
2010-11-07T01:17:08  <amstan> McLeopold: here's the corner stuff: http://csclub.uwaterloo.ca/~amstan/upload/sauer/terrain/
2010-11-07T01:19:00  <janzert> C# compilation on the main site shouldn't be throwing unexpected errors anymore since a week or so ago
2010-11-07T01:19:15  <McLeopold> I think my fault line assumption was correct.
2010-11-07T01:19:36  <amstan> McLeopold: yeah.. the look in the 8 corners around
2010-11-07T01:20:43  <McLeopold> but, you're gonna have issues with large maps and recursion, unless you do tail
2010-11-07T01:21:20  <amstan> McLeopold: yeah... i'm already having huge issues, here's the code that converts an image to the that map that you saw: https://github.com/amstan/hyperserv/blob/master/src/pyscripts/editing/image.py
2010-11-07T01:21:25  <amstan> specifically loadimage2
2010-11-07T01:23:29  <McLeopold> I wonder if you could do a top down, or bottom up approach
2010-11-07T01:23:55  <McLeopold> Order the grids by height, then fill those in first.
2010-11-07T01:24:15  <McLeopold> Always heal toward the already established column.
2010-11-07T01:24:42  <amstan> thing is that you cannot tell if there's a cube there already or not
2010-11-07T01:24:55  <amstan> i'm hacking this onto the servers, forging packets
2010-11-07T01:25:04  <McLeopold> oh
2010-11-07T01:25:06  <amstan> the server's not aware of the map contents actually
2010-11-07T01:25:16  <amstan> it just blindly forwards packets
2010-11-07T01:25:17  <McLeopold> can you get the contents first?
2010-11-07T01:25:35  <McLeopold> Actually, that implementation is beyond what I want to understand. :)
2010-11-07T01:26:26  <amstan> thing is.. that the image is all you have, you cannot tell if you built a column somewhere or not, unless you specifically store it in another array
2010-11-07T01:27:21  <Prillicy> wow, that had a lot of bugs
2010-11-07T01:27:24  <Prillicy> I must be tired
2010-11-07T01:27:46  <McLeopold> yeah! more good sniping. http://72.44.46.68/canvas?game_id=889961
2010-11-07T01:28:09  <Prillicy> contestbot help
2010-11-07T01:28:09  <contestbot> Prillicy: (help [<plugin>] [<command>]) -- This command gives a useful description of what <command> does. <plugin> is only necessary if the command is in more than one plugin.
2010-11-07T01:28:20  <Prillicy> help all
2010-11-07T01:28:26  <Prillicy> contestbot: help all
2010-11-07T01:28:26  <contestbot> Prillicy: Error: There is no command "all".
2010-11-07T01:28:31  <Prillicy> contestbot: help commands
2010-11-07T01:28:31  <contestbot> Prillicy: (commands takes no arguments) -- Returns a list of the commands offered by the bot.
2010-11-07T01:28:41  <Prillicy> contestbot: commands
2010-11-07T01:28:41  <contestbot> Prillicy: action, add, alert, alias, allyourspace, announce, announce add, announce list, announce remove, any, aol, apply, apropos, at, author, azn, ban add, ban list, ban remove, base, binary, bold, bye, cache, calc, call, capabilities, capability add, capability list, capability remove, capability set, capability setdefault, capability unset, capitalize, change, changename, channel, channels, (5 more messages)
2010-11-07T01:28:42  <amstan> Prillicy: tip, do commands in a pm
2010-11-07T01:28:45  <McLeopold> http://ubottu.com/stdin/supydocs/commands.html
2010-11-07T01:28:58  <amstan> or that..
2010-11-07T01:29:20  <Prillicy> I honestly just want to see where I am in the rankings before I upload a new bot (which I need to get to compile first
2010-11-07T01:29:33  <amstan> Prillicy: no function for that
2010-11-07T01:29:42  <Prillicy> hmm
2010-11-07T01:29:42  <McLeopold> @ranking 10000
2010-11-07T01:29:42  <contestbot> McLeopold: Error: Cannot do that many players.
2010-11-07T01:29:47  <McLeopold> :(
2010-11-07T01:29:53  <Prillicy> aww
2010-11-07T01:30:15  <McLeopold> @rankings 10-20
2010-11-07T01:30:15  <contestbot> McLeopold: (rankings [how many]) -- Get the top $(how many) players.
2010-11-07T01:30:40  <amstan> @botrepo
2010-11-07T01:30:41  <contestbot> amstan: botrepo = http://github.com/amstan/contestbot
2010-11-07T01:30:42  <McLeopold> Prillicy: have you done tcp yet?
2010-11-07T01:30:47  <amstan> feel free to add moar
2010-11-07T01:31:00  <Prillicy> yes, a little
2010-11-07T01:31:03  <Prillicy> it'ss confusing
2010-11-07T01:31:21  <McLeopold> Try the JBotManager by delt0r.  I thinks its great.
2010-11-07T01:31:27  <McLeopold> You can watch tcp matches live.
2010-11-07T01:31:30  <Prillicy> I can't see it visually in that
2010-11-07T01:31:34  <Prillicy> but it seems to show up in my console
2010-11-07T01:31:43  <Prillicy> still, the problem of compiling remains
2010-11-07T01:32:01  <amstan> Prillicy: did you try compiling with mono?
2010-11-07T01:32:12  <Prillicy> mono?
2010-11-07T01:32:36  <amstan> since c# was not designed for running on linux initially, someone had to write an implementation for it
2010-11-07T01:32:40  <amstan> it's not perfect..
2010-11-07T01:33:37  <Prillicy> I don't have linux
2010-11-07T01:34:04  <McLeopold> Prillicy: try installing VirtualBox and then installing a linux distro inside.
2010-11-07T01:34:22  <amstan> Prillicy: or ask someone who has linux and mono experience to do it for you
2010-11-07T01:34:26  <Prillicy> just so I can run mono to compile c# outside of its native environment?
2010-11-07T01:34:40  <McLeopold> Well, this is a learning opportunity.
2010-11-07T01:34:42  <Prillicy> amstan, you want to do me a favor?
2010-11-07T01:34:52  <amstan> Prillicy:  i don't have mono experience
2010-11-07T01:35:01  <Prillicy> McLeopold: which  will do me no good.  I'm learning c# for work, where we use windows
2010-11-07T01:35:02  <amstan> nor do i know how to get the exact version from the server
2010-11-07T01:35:19  <amstan> Prillicy: don't discard that as no good
2010-11-07T01:35:54  <Prillicy> I'm not going to get into an OS debate here
2010-11-07T01:36:00  <Prillicy> but I will not install linux
2010-11-07T01:36:14  <janzert> Mono JIT compiler version 2.4.4
2010-11-07T01:36:16  <amstan> if you're the only one at work who knows how to port c# stuff to other platforms it'll definitelly help
2010-11-07T01:36:28  <amstan> janzert: is that easily gettable on ubuntu 10.10?
2010-11-07T01:36:38  <janzert> no idea
2010-11-07T01:36:43  <amstan> :(
2010-11-07T01:36:44  <amstan> k
2010-11-07T01:36:55  <amstan> it's a start.. knowing the version
2010-11-07T01:37:06  <amstan> janzert: do you know how to compile?
2010-11-07T01:37:18  <janzert> it seems to be the newest version easily gettable on 8.04 though :/
2010-11-07T01:37:34  <janzert> mono programs?
2010-11-07T01:37:38  <amstan> i can get 2.6
2010-11-07T01:38:03  <janzert> not off the top of my head, the method used for the contest is in compile_anything of course
2010-11-07T01:38:16  <amstan> #repo
2010-11-07T01:38:18  <amstan> @repo
2010-11-07T01:38:19  <contestbot> amstan: repo = http://code.google.com/p/ai-contest/
2010-11-07T01:38:23  <amstan> Prillicy: k... i'll do this for you
2010-11-07T01:38:32  <Prillicy> ?
2010-11-07T01:38:39  <amstan> Prillicy: but don't refer me to others, i only help like this once in a blue moon
2010-11-07T01:38:48  <Prillicy> thank you so much
2010-11-07T01:38:51  <Prillicy> you're the best
2010-11-07T01:38:53  <Prillicy> do you want the code?
2010-11-07T01:39:05  <amstan> Prillicy: i'm also assuming your code won't make my comp be part of a bot network
2010-11-07T01:39:30  <Prillicy> I was going to pastebin it for you to copy paste to a .cs file
2010-11-07T01:39:38  <Prillicy> can't get much more obvious than that
2010-11-07T01:39:49  <amstan> isn't this it? http://pastebin.com/3uAM81Ex
2010-11-07T01:39:51  <janzert> gmcs -warn:0 -out:MyBot.exe *.cs is basically the command used by compile_anything
2010-11-07T01:40:22  <Prillicy> http://pastebin.com/6RMJ4XFY <--fixed some bugs
2010-11-07T01:40:35  <janzert> except that it actually lists out each .cs file
2010-11-07T01:41:08  <amstan> downloading compiler
2010-11-07T01:41:14  <amstan> installing..
2010-11-07T01:41:45  <amstan> bot.cs(146,31): error CS0246: The type or namespace name `Planet' could not be found. Are you missing a using directive or an assembly reference?
2010-11-07T01:41:54  <amstan> i'm assuming i need some external library file
2010-11-07T01:42:01  <amstan> Prillicy: i can has that pls?
2010-11-07T01:42:12  <amstan> Prillicy: i hope you're including that in the zip though
2010-11-07T01:42:22  <Prillicy> oh, yeah, the 3 other files are in the starter package
2010-11-07T01:42:34  <Prillicy> and of course, but I need to get it to compile here before I upload it
2010-11-07T01:42:46  <Prillicy> if you want them, I can pastebin them also
2010-11-07T01:43:26  <amstan> 29 errors
2010-11-07T01:43:38  <amstan> Prillicy: http://pastebin.com/JBHhvvj8
2010-11-07T01:44:06  <Prillicy> thanks a million
2010-11-07T01:44:11  <amstan> trying to access private members?
2010-11-07T01:44:34  <amstan> janzert: so why won't the emails work then?
2010-11-07T01:44:54  <janzert> they work for me every time
2010-11-07T01:44:58  <amstan> oh, lol
2010-11-07T01:45:20  <janzert> My thought is that maybe some ISP's are dropping them as spam
2010-11-07T01:45:26  <janzert> other than that I don't really know
2010-11-07T01:45:35  <amstan> janzert: also, did you think about compiling python?
2010-11-07T01:45:56  <McLeopold> http://www.codinghorror.com/blog/2010/04/so-youd-like-to-send-some-email-through-code.html
2010-11-07T01:46:01  <janzert> mostly thought about adding a simple test case everything would get run through
2010-11-07T01:46:08  <janzert> McLeopold: yeah
2010-11-07T01:46:41  <janzert> amstan: but it's not quite trivial so I'm not actively working on it right now
2010-11-07T01:46:53  * Prillicy hugs amstan
2010-11-07T01:47:05  <amstan> janzert: isn't it?
2010-11-07T01:47:23  <janzert> not quite
2010-11-07T01:47:32  <amstan> janzert: can i help somehow?
2010-11-07T01:47:43  <amstan> i can ask around in #python, try it myself
2010-11-07T01:47:55  <janzert> it really wouldn't be hard but I'm working on some ranking and game matchup stuff right now
2010-11-07T01:48:06  <McLeopold> i think it is, compiling python
2010-11-07T01:48:08  <amstan> ok, cool
2010-11-07T01:48:21  <amstan> McLeopold: i mean the stuff from py->pyc
2010-11-07T01:48:37  <janzert> oh you're talking about compiling python, I meant running every new submission through a simple testcase
2010-11-07T01:49:08  <McLeopold> import py_compile
2010-11-07T01:49:29  <McLeopold> janzert: I know some bots expect exactly 200 turns and 23 planets or they break.
2010-11-07T01:49:29  <amstan> janzert: yes... it's a small speed increase in the first turn, and people will be able to receive syntax errors by mail on that upload step
2010-11-07T01:49:46  <McLeopold> very small
2010-11-07T01:50:01  <amstan> McLeopold: but we do this for other languages though
2010-11-07T01:50:02  <McLeopold> but the syntax thing is a great upside
2010-11-07T01:50:20  <janzert> McLeopold: it the test case would basically simulate a first turn and then kill the bot
2010-11-07T01:50:41  <amstan> janzert: we can have that for next contest
2010-11-07T01:50:49  <janzert> as long as the bot returned a valid first turn it would pass
2010-11-07T01:51:09  <amstan> no man! it would have to win!
2010-11-07T01:51:14  <McLeopold> create a dumb bot that wait 1 turn, then give an invalid.  Make that everyone's first match.
2010-11-07T01:51:18  <amstan> like.. if it lost in the first turn.. just don't even bother, lol
2010-11-07T01:51:27  <amstan> not talking about invalid moves here though
2010-11-07T01:51:52  <janzert> you would catch syntax errors in all the non-compiled languages that way
2010-11-07T01:52:06  <amstan> janzert: runtime too
2010-11-07T01:52:07  <janzert> as well as a whole pile of other bugs of course
2010-11-07T01:52:35  <amstan> it's an interesting idea..
2010-11-07T01:53:01  <amstan> but by now.. i don't think we'll have many more new people that would need this joining
2010-11-07T01:53:48  <janzert> about 30-50 a day it seems and really I think it's not just new people that would benefit a lot
2010-11-07T01:54:38  <amstan> 30-50? huh?
2010-11-07T01:54:44  <janzert> http://ai-contest.com/game_stats.php
2010-11-07T01:55:02  <janzert> it seems to be the lowest over the weekend and sort of builds through the week
2010-11-07T01:57:15  *** Itkovian <Itkovian!~Itkovian@2002:4e16:ea86:0:21f:5bff:fec8:6e2c> has joined #aichallenge
2010-11-07T02:01:37  <amstan> woah. how did we get to 600 games per minute?
2010-11-07T02:01:56  <janzert> what?
2010-11-07T02:02:03  <amstan> http://ai-contest.com/game_stats.php
2010-11-07T02:02:08  <janzert> more like 30 to 50
2010-11-07T02:02:14  <McLeopold> lots of crashes?
2010-11-07T02:02:18  <amstan> last 5 min: 588gpm, last hour: 80gpm
2010-11-07T02:02:22  <janzert> woah
2010-11-07T02:02:34  <janzert> only 10%
2010-11-07T02:02:59  <janzert> oh, haha
2010-11-07T02:03:02  <amstan> all the servers at once too
2010-11-07T02:03:02  <janzert> time change
2010-11-07T02:03:07  <amstan> oh!
2010-11-07T02:03:07  <janzert> the averages are messed up
2010-11-07T02:03:08  <amstan> lol
2010-11-07T02:03:11  <McLeopold> haha
2010-11-07T02:03:22  <McLeopold> take a screen shot a brag on the forum
2010-11-07T02:03:26  <janzert> had me scared there
2010-11-07T02:03:57  <McLeopold> I wonder how much code break twice a year due to that.
2010-11-07T02:04:29  <amstan> McLeopold: apparently it costs more to maintain daylightsavings than to actually use lights without daylight saving
2010-11-07T02:05:01  <McLeopold> unless it's just year round.
2010-11-07T02:05:25  <McLeopold> if all of humanity didn't sleep in so much, it wouldn't be an issue
2010-11-07T02:05:50  <amstan> we should all do polyphasic sleep
2010-11-07T02:06:24  <McLeopold> siesta's would be healthier
2010-11-07T02:07:15  <amstan> McLeopold: does that shorten the sleep time per day?
2010-11-07T02:07:35  <McLeopold> No
2010-11-07T02:07:51  <McLeopold> We need a good 8 hours of sleep a day.
2010-11-07T02:08:09  <janzert> in this case it's actually mysql's interval calculation code that is causing the bug so I bet it bites a lot of code out there
2010-11-07T02:08:40  <amstan> why is mysql not using unix time?
2010-11-07T02:08:44  <McLeopold> doesn't unix just slow time down until it is synced again?
2010-11-07T02:08:46  <amstan> it would not be an issue then
2010-11-07T02:08:54  <janzert> McLeopold: nope
2010-11-07T02:09:08  <janzert> amstan: no idea
2010-11-07T02:11:01  <McLeopold> because of leap seconds?
2010-11-07T02:13:06  *** Itkovian <Itkovian!~Itkovian@2002:4e16:ea86:0:21f:5bff:fec8:6e2c> has quit IRC (Quit: Itkovian)
2010-11-07T02:13:53  <janzert> daylight savings time is a purely 'local' convention and the adjustments are made as specified basically
2010-11-07T02:14:15  <amstan> a1k0n: so.. what's going on? not wanting to own everyone?
2010-11-07T02:14:20  <amstan> once again?
2010-11-07T02:15:36  <contestbot> Top 10 players: bocsimacko(4322), shangas(4174), Raschi(4118), bix0r4ever(4093), GreenTea(4054), rsergio(4049), Xrillo(4035), george(4032), dmj111(4024), Slin-.-(4021)
2010-11-07T02:18:23  <amstan> i kinda like this competition more than tron, since other languages have a chance besides C/C++
2010-11-07T02:18:38  <amstan> minimax spinning is not essential anymore
2010-11-07T02:21:12  <McLeopold> not yet
2010-11-07T02:22:05  *** tapwater <tapwater!~tapwater@cpe-70-112-195-55.austin.res.rr.com> has quit IRC (Quit: tapwater)
2010-11-07T02:22:21  <raedwulf> if someone figures minimax hehe
2010-11-07T02:22:31  <raedwulf> it'll probably have to be a compromise
2010-11-07T02:23:27  <sigh> well, I think many people are *trying* it
2010-11-07T02:23:32  <sigh> I personally had no luck with it
2010-11-07T02:23:40  <raedwulf> hmmm, and the top bots aren't minimax
2010-11-07T02:23:53  <sigh> raedwulf: how do you knw?
2010-11-07T02:24:01  <raedwulf> well mega1 isn't
2010-11-07T02:24:10  <raedwulf> - he said so :P
2010-11-07T02:24:12  <sigh> :)
2010-11-07T02:24:34  <sigh> I don't think dmj is either
2010-11-07T02:24:59  <raedwulf> as i said, it wouldn't be a pure minimax, and useful hybrids are probably very hard to do
2010-11-07T02:25:17  <sigh> yeah, well pure minimax is our of the question, that's a given
2010-11-07T02:25:18  *** Rubicon-|-Cross <Rubicon-|-Cross!~RubiconCr@adsl-99-30-224-80.dsl.pltn13.sbcglobal.net> has joined #aichallenge
2010-11-07T02:25:30  <raedwulf> because you no doubt get diminishing returns as soon as you try and prune the tree
2010-11-07T02:25:53  <raedwulf> because if you look at what you *think* are good, you're going all heuristic again
2010-11-07T02:25:59  <McLeopold> yeah, but dimished from perfect may still be superiour to heuristic
2010-11-07T02:26:18  <raedwulf> thats the question tho - its tricky to say
2010-11-07T02:26:21  <McLeopold> minimix still relies on an eval function
2010-11-07T02:26:27  <sigh> yup
2010-11-07T02:26:44  <McLeopold> and a move generator with good ordering
2010-11-07T02:26:47  <raedwulf> but the more complex your eval function, the less you can compute in good enough time
2010-11-07T02:27:13  <sigh> I don't think the eval function needs to be very complex for this game
2010-11-07T02:27:30  <sigh> it's the pruning that requires complex smarts
2010-11-07T02:28:04  <antimatroid> i disagree
2010-11-07T02:28:35  <amstan> Prillicy: compiled?
2010-11-07T02:28:46  <antimatroid> i think the "right" move changes so quickly that it's important to have a really good evaluation functions, plus you can't hope to search anything near the breadth of the real game tree to even prune
2010-11-07T02:28:58  <Prillicy> almost, having a little trouble passing the planetwars class to my array comparison method
2010-11-07T02:29:04  <Prillicy> tbh, I have no clue how
2010-11-07T02:30:37  <Prillicy> it needs the pw class to call some of its functions
2010-11-07T02:31:22  <antimatroid> raedwulf: mega1 said he was doing minimax sort of at one point
2010-11-07T02:35:45  <raedwulf> antimatroid: ah
2010-11-07T02:36:16  <raedwulf> i still have my idea to test on minimax
2010-11-07T02:36:28  <raedwulf> but damn, i've got barely enough time ot get this one to work :P
2010-11-07T02:37:18  *** yasith <yasith!~yasith@112.135.65.77> has quit IRC (Read error: Connection reset by peer)
2010-11-07T02:42:34  <antimatroid> i have a bunch of ways to possibly do it, it all depends on me resolving battles into the future properly though
2010-11-07T02:42:47  <antimatroid> i have exams tomorrow and tuesday, then i think i might do a rewrite and try to get it right
2010-11-07T02:43:40  *** Azrathud <Azrathud!~azrathud@c-67-166-43-227.hsd1.co.comcast.net> has quit IRC (Quit: Meh)
2010-11-07T02:49:38  <Prillicy> is it ok to instantiate a new planetwars class?
2010-11-07T02:49:56  *** yasith <yasith!~yasith@112.135.109.15> has joined #aichallenge
2010-11-07T02:50:38  *** jaspervdj <jaspervdj!~jaspervdj@zeus.ugent.be> has joined #aichallenge
2010-11-07T02:51:22  <Prillicy> bah, I can't access planetwars methods because the class needs to be instantiated, but if I do, I have to gather info from the server which creates other problems
2010-11-07T02:52:47  <raedwulf> Prillicy: you can modify the starter pack code -its just there for a sample
2010-11-07T02:53:02  <raedwulf> some people rewrite their code completely from scratch
2010-11-07T02:53:06  <Prillicy> yeah, I know, but I also don'treally know what I'm doing
2010-11-07T02:53:14  <raedwulf> :(
2010-11-07T02:54:47  <Prillicy> well, instantiating a new PW class crashed the bot
2010-11-07T02:54:48  <Prillicy> yay
2010-11-07T02:55:00  <Prillicy> so I need a better way to access these functions
2010-11-07T02:55:20  <Prillicy> I think I'm going to start by putting them all in one file (I'm guessing that's bad practice?)
2010-11-07T03:02:19  <raedwulf> probably
2010-11-07T03:02:40  <janzert> if c# is like other starter packs your doturn function should be getting handed a PlanetWars object
2010-11-07T03:04:46  *** McLeopold <McLeopold!~scott@c-76-121-152-123.hsd1.wa.comcast.net> has quit IRC (Quit: Leaving)
2010-11-07T03:10:35  <Prillicy> janzert: it is, but my array comparison doesn't...not a worry, I found a way to make it work
2010-11-07T03:15:37  <contestbot> Top 10 players: bocsimacko(4372), shangas(4230), Raschi(4145), bix0r4ever(4119), george(4087), rsergio(4085), Xrillo(4063), GreenTea(4062), dmj111(4053), Slin-.-(4038)
2010-11-07T03:20:12  <Prillicy> I'm getting illegal fleet request errors left and rght
2010-11-07T03:20:33  <Prillicy> is that common?
2010-11-07T03:22:59  <raedwulf> only if your code is wrong
2010-11-07T03:24:57  <sigh> having wrong code is incredibly common
2010-11-07T03:26:14  <raedwulf> :)
2010-11-07T03:33:00  <Prillicy> hmm, now I've got an IO exception
2010-11-07T03:33:07  <Prillicy> does anyone have the energy to take a look
2010-11-07T03:33:09  <Prillicy> ?
2010-11-07T03:38:03  <Prillicy> gues not
2010-11-07T03:42:08  <nullkuhl> Prillicy: pastebin ?
2010-11-07T03:42:52  <Prillicy> http://pastebin.com/qdF3Hiqp <---the whole thing
2010-11-07T03:43:23  <nullkuhl> where does the IO exception happen
2010-11-07T03:43:39  *** adhoc <adhoc!~adhoc@ppp121-45-60-60.lns20.adl2.internode.on.net> has joined #aichallenge
2010-11-07T03:46:26  <nullkuhl> Prillicy: ?
2010-11-07T03:46:35  <Prillicy> http://area-51-x.com/css/images/java-errors.png
2010-11-07T03:48:11  <Prillicy> sorry nullkuhl
2010-11-07T03:48:21  <Prillicy> took me a minute to host it
2010-11-07T03:53:57  *** Appleman1234 <Appleman1234!~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au> has quit IRC (Read error: Connection reset by peer)
2010-11-07T04:04:11  <bhasker> its weird when i add more code to account for future growth/ships etc my bot gets worse:-\
2010-11-07T04:04:23  <bhasker> giving it better state information seems to paralyze the bot
2010-11-07T04:04:30  <bhasker> especially when it starts losing
2010-11-07T04:04:31  <bhasker> :-\
2010-11-07T04:06:20  <raedwulf> :(
2010-11-07T04:08:10  <Prillicy> it gets scared?
2010-11-07T04:09:17  <Prillicy> my bot can't get through one game without crashing if it makes ya feel better
2010-11-07T04:12:03  *** Itkovian <Itkovian!~Itkovian@c083h027.ugent.be> has joined #aichallenge
2010-11-07T04:12:23  <bhasker> it just figures that enemy will have more ships etc and thinks it can't send out any ships without losing
2010-11-07T04:12:34  <bhasker> so it just kind of grinds to a hal
2010-11-07T04:12:35  <bhasker> t
2010-11-07T04:13:03  <Prillicy> lol
2010-11-07T04:13:30  <sigh> bhasker: yeah, that's the danger with a lot of techniques
2010-11-07T04:13:40  <sigh> pretty much the same issue comes up in minimax
2010-11-07T04:13:57  <bhasker> not sure how i can work around that
2010-11-07T04:14:10  <bhasker> basically if i even go down in growth by a couple of points
2010-11-07T04:14:15  <bhasker> its game over
2010-11-07T04:14:26  <bhasker> since it figures every planet can't send out any ships
2010-11-07T04:14:29  <bhasker> it stops staging
2010-11-07T04:14:36  <bhasker> then if enemy attacks my outermost
2010-11-07T04:14:38  <bhasker> its gg
2010-11-07T04:14:46  <bhasker> he takes it before my reinforcements can arrive
2010-11-07T04:15:06  <bhasker> since once he commits his ships to flight my bot will react
2010-11-07T04:15:31  <sigh> yeah, one way I get around that is by purposly overriding defence in various situations
2010-11-07T04:15:36  <contestbot> Top 10 players: bocsimacko(4368), shangas(4218), Raschi(4134), bix0r4ever(4126), rsergio(4096), george(4086), Xrillo(4061), GreenTea(4052), dmj111(4035), Hazard(4020)
2010-11-07T04:15:48  <bhasker> what do you mean overriding defense
2010-11-07T04:15:54  <bhasker> you send ships to front anyway?
2010-11-07T04:16:00  <bhasker> as long as the planet is not under attack?
2010-11-07T04:16:53  <sigh> for example one case is that, if planet A is keeping planets in case planet B attacks... then I allow that to be overriden if planet A wants to attack planet B
2010-11-07T04:17:29  <bhasker> hmm thats one thing i haven't yet figured out how to do
2010-11-07T04:17:40  <bhasker> keeping track of which planets affect which planet
2010-11-07T04:18:14  <sigh> I purposly keep a very simple model, so I can do exclusions like that
2010-11-07T04:19:12  *** shades <shades!~shades@cheshire.telephreak.org> has quit IRC (Ping timeout: 264 seconds)
2010-11-07T04:19:19  <sigh> another thing is overriding defence if the planet A wants to attack a planet with a higher growth rate... but I think I had some prblems with that
2010-11-07T04:22:30  <bhasker> hmm
2010-11-07T04:22:42  <bhasker> oh well
2010-11-07T04:23:00  <bhasker> i need to figure out
2010-11-07T04:23:01  <bhasker> http://72.44.46.68/canvas?game_id=892241
2010-11-07T04:23:13  <bhasker> why my bot didnt take the 5 growth with 1 ship left from the two nearby planets
2010-11-07T04:23:25  <bhasker> and waited till the 2 ships from half way across the map reached it
2010-11-07T04:29:43  *** Appleman1234 <Appleman1234!~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au> has joined #aichallenge
2010-11-07T04:31:45  <bhasker> damn coordinated attacks
2010-11-07T04:31:59  <bhasker> stupid thing calculated that a coordinated attack was in progress and didnt launch
2010-11-07T04:32:17  <bhasker> i should add a check to override it if the order > numships doh
2010-11-07T04:43:14  *** Appleman1234 <Appleman1234!~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au> has quit IRC (Ping timeout: 260 seconds)
2010-11-07T04:46:04  *** boegel <boegel!~boegel@vpnq243.ugent.be> has joined #aichallenge
2010-11-07T04:46:28  *** Itkovian <Itkovian!~Itkovian@c083h027.ugent.be> has left #aichallenge
2010-11-07T04:49:38  <bhasker> JBotManager is awesome
2010-11-07T04:50:43  <antimatroid> i don't maintain coordinated attacks between moves
2010-11-07T04:50:50  <antimatroid> if they're still optimal, they'll be picked up
2010-11-07T04:52:05  <sigh> yeah, saving state just makes things more complicated
2010-11-07T04:55:48  <nullkuhl> bhasker: what does JBotManager mainly do ?
2010-11-07T04:56:05  <bhasker> it lets you watch games live
2010-11-07T04:56:19  <bhasker> i don't save state either
2010-11-07T04:56:40  <bhasker> its just that my coordinating logic will prevent new launches while other fleets are on the way
2010-11-07T04:56:46  *** Appleman1234 <Appleman1234!~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au> has joined #aichallenge
2010-11-07T04:56:47  <bhasker> so that they launch synchronizedd
2010-11-07T04:56:58  <bhasker> but i guess the code was broken
2010-11-07T04:57:55  <nullkuhl> bhasker: would you kindly post a snapshot of what does it display or so ?
2010-11-07T04:58:09  <bhasker> hmm let me pull up the forum post for it
2010-11-07T04:58:57  <bhasker> http://ai-contest.com/forum/viewtopic.php?f=18&t=751
2010-11-07T04:59:00  <bhasker> just run it and see
2010-11-07T04:59:10  *** shades <shades!~shades@cheshire.telephreak.org> has joined #aichallenge
2010-11-07T05:00:08  <nullkuhl> bhasker: its that i have trouble with tcp already, i just wanted how it looks like to know if its worth troubleshooting and fixing
2010-11-07T05:00:34  <bhasker> its just a java program
2010-11-07T05:00:43  <bhasker> run it and see , that post describes what it does
2010-11-07T05:00:50  <nullkuhl> the name is enough to tell that :), nvm anyway
2010-11-07T05:02:35  <bhasker> it will just run your bot
2010-11-07T05:02:40  <bhasker> connect to tcp server
2010-11-07T05:02:44  <bhasker> and let you watch the game live
2010-11-07T05:02:50  <bhasker> it also allows you to save every game
2010-11-07T05:02:51  <bhasker> etc
2010-11-07T05:04:47  <nullkuhl> does it has its own visualizer ? or uses that java one that comes in starter packages
2010-11-07T05:05:14  <bhasker> it has its own visualizer
2010-11-07T05:05:16  <bhasker> its pretty neat
2010-11-07T05:05:23  <bhasker> shows you lines/graphs of growth etc
2010-11-07T05:06:54  <nullkuhl> hmm just saw it
2010-11-07T05:07:17  <nullkuhl> i think i can improve various stuff here, as a java dev
2010-11-07T05:07:20  <nullkuhl> is it open source ?
2010-11-07T05:15:38  <contestbot> Top 10 players: bocsimacko(4367), shangas(4220), Raschi(4125), bix0r4ever(4109), rsergio(4087), george(4079), Xrillo(4051), asavis(4044), GreenTea(4044), dmj111(4034)
2010-11-07T05:20:49  <delt0r_> nullkuhl, JBotManager is open source... Its included in the jar
2010-11-07T05:21:12  <delt0r_> I accept patches too if you want to make your changes public
2010-11-07T05:32:12  *** irchs <irchs!~Jan@2002:5ec2:21a8:0:226:bbff:fe1d:f630> has joined #aichallenge
2010-11-07T05:32:14  <nullkuhl> delt0r_: awesome
2010-11-07T05:37:22  <lavalamp_> I made contest bot learn about jbotmanager
2010-11-07T05:37:28  <lavalamp_> @jbotmanager
2010-11-07T05:37:29  <contestbot> lavalamp_: jbotmanager = http://www.mabs.at/ewing/JBotManager.jar
2010-11-07T05:44:54  <nullkuhl> i couldnt beat my own bot in Human mode in jbotmanager XD
2010-11-07T05:46:47  *** Migi32 <Migi32!~migi@d54C1291D.access.telenet.be> has joined #aichallenge
2010-11-07T06:02:53  <antimatroid> i can't believe there isn't a wikipedia page on integral systems
2010-11-07T06:03:02  <antimatroid> wikipedia is starting to fail me these days
2010-11-07T06:03:56  <antimatroid> i can beat my bot in jbotmanager depending on the map
2010-11-07T06:04:13  <bhasker> i should try that
2010-11-07T06:04:26  <bhasker> but on complex maps it would get quite hard for a human to gauge stuff
2010-11-07T06:04:39  <antimatroid> yeah, i need to fuck it up in the first few moves
2010-11-07T06:05:18  <bhasker> well i have got my bot to a point where it doesn't give u
2010-11-07T06:05:21  <bhasker> up*
2010-11-07T06:05:23  <bhasker> that quickly
2010-11-07T06:05:30  <bhasker> but still has issues
2010-11-07T06:05:31  <bhasker> sigh
2010-11-07T06:05:37  <bhasker> not @sigh:p
2010-11-07T06:05:46  <sigh> :P
2010-11-07T06:08:23  <lavalamp_> is there a command line engine that runs faster than the stock playgame.jar?
2010-11-07T06:09:39  <bhasker> get the c++ version
2010-11-07T06:09:42  <bhasker> its much faster
2010-11-07T06:09:50  <bhasker> its there on the forums somewhere
2010-11-07T06:13:23  <lavalamp_> hm what's it called?
2010-11-07T06:14:02  *** fawek <fawek!~faw@chello089079023212.chello.pl> has joined #aichallenge
2010-11-07T06:15:10  *** Commandir <Commandir!~joakim@h85-8-40-32.static.se.alltele.net> has joined #aichallenge
2010-11-07T06:15:37  <contestbot> Top 10 players: bocsimacko(4394), shangas(4237), Raschi(4128), bix0r4ever(4118), george(4099), rsergio(4088), dmj111(4056), Xrillo(4054), asavis(4049), GreenTea(4043)
2010-11-07T06:17:41  <nullkuhl> lavalamp_: i was searching for the same thing, not sure if this is it : http://ai-contest.com/forum/viewtopic.php?f=18&t=550&p=3775&hilit=playgame+c#p3775
2010-11-07T06:18:27  <lavalamp_> yeah thanks I just found it
2010-11-07T06:26:10  *** mceier <mceier!~mceier@chello089077208118.chello.pl> has joined #aichallenge
2010-11-07T06:27:29  <antimatroid> sigh sort of has to expect that with a name like that
2010-11-07T06:27:42  <antimatroid> le sigh :P
2010-11-07T06:27:58  <antimatroid> le sighered
2010-11-07T06:28:08  <sigh> I've found that people in this channel don't sigh that much
2010-11-07T06:28:32  <sigh> it's mostly bhasker, actually :P
2010-11-07T06:28:32  <antimatroid> i don't think i ever do textually
2010-11-07T06:28:49  <bhasker> :p
2010-11-07T06:28:57  <lavalamp_> I keep wanting to sigh and then realizing that there's a sigh here XP
2010-11-07T06:28:59  <antimatroid> i did quite a lot in calculus of variations last semester
2010-11-07T06:29:09  <antimatroid> some women who didn't make it through semester made some really strange comments
2010-11-07T06:29:17  <antimatroid> the worst being to ask if a 1 should have ^2 next to it
2010-11-07T06:29:29  <lavalamp_> :D
2010-11-07T06:29:33  <sigh> maybe it was a large value f 1
2010-11-07T06:29:35  <sigh> *of
2010-11-07T06:29:55  <antimatroid> seriously, i don't know how one even gets in to a class like that and asks such a question
2010-11-07T06:29:58  <lavalamp_> haha wow the c++ one really is a lot faster
2010-11-07T06:30:16  * sigh aims to get into a class like that and ask that question
2010-11-07T06:30:26  <lavalamp_> I don't know how you write an engine slower than the bots...
2010-11-07T06:33:42  <delt0r_> lavalamp_, just about every engine thats not playgame.jar is much faster
2010-11-07T06:33:57  <lavalamp_> yeah that's crazy
2010-11-07T06:34:01  <delt0r_> well
2010-11-07T06:34:08  <lavalamp_> jbotmanager is great, btw
2010-11-07T06:34:14  <lavalamp_> thanks for making it
2010-11-07T06:34:21  <delt0r_> Its what happens when a C programer programs java....
2010-11-07T06:34:27  <lavalamp_> haha
2010-11-07T06:34:34  <delt0r_> which is still better than a java programmer programming C
2010-11-07T06:34:41  <delt0r_> thx, np
2010-11-07T06:35:20  <antimatroid> delt0r_: i have but one complaint, why can't i specify exact ship numbers to send when playing as a human?
2010-11-07T06:35:25  <antimatroid> otherwise <3<3<3
2010-11-07T06:35:54  <delt0r_> antimatroid, Because I don't know how to put that into an interface properly
2010-11-07T06:36:01  <delt0r_> I sux at GUI
2010-11-07T06:36:06  <delt0r_> and try to make it simple
2010-11-07T06:36:09  <sigh> :)
2010-11-07T06:36:15  <delt0r_> suggestions welcome
2010-11-07T06:36:21  <sigh> yeah, it's a tricky problem
2010-11-07T06:37:40  *** p4p4 <p4p4!~chatzilla@82.113.121.50> has joined #aichallenge
2010-11-07T06:37:52  <antimatroid> i have never done anything gui related, my best guess would be to do something like continue changing the variable currently doing proportions every time a number is hit, reseting it to 0 every time they change planet and incrementing multipler by *10 every time they hit a number of ships to send
2010-11-07T06:37:55  <antimatroid> if that makes sense at all
2010-11-07T06:38:22  <delt0r_> not really :/
2010-11-07T06:39:35  <antimatroid> let n = ships to send, m = multiplier
2010-11-07T06:39:36  <antimatroid> player highlights planet -> n= 0, m=1
2010-11-07T06:39:36  <antimatroid> player hits number -> n+= m*numberHit, m*=10
2010-11-07T06:39:36  <antimatroid> player hits enter, send n ships
2010-11-07T06:39:58  <antimatroid> wait, no
2010-11-07T06:40:04  <antimatroid> those numbers need to be reversed
2010-11-07T06:40:06  <antimatroid> if that makes sense
2010-11-07T06:40:35  <delt0r_> mmm
2010-11-07T06:40:39  <delt0r_> yea i get it now
2010-11-07T06:40:40  <sigh> that's the mostly complicated way of explaining how someone would enter a number as a string
2010-11-07T06:40:45  <sigh> if I understand correctly
2010-11-07T06:40:56  <antimatroid> haha sorry, that's what i could think of
2010-11-07T06:41:09  <antimatroid> i was thinking in numbers
2010-11-07T06:41:14  <sigh> :P
2010-11-07T06:41:15  <delt0r_> don't think it would really work in practice... what happens when you have N planets selected for example?
2010-11-07T06:41:22  <antimatroid> one could just input a string and convert it to a number upon sending hte fleet
2010-11-07T06:41:40  <delt0r_> that could be done...
2010-11-07T06:41:42  <antimatroid> i'd rather send indivually like this and up the time limit
2010-11-07T06:42:01  <delt0r_> perhaps the number pad does string numbers
2010-11-07T06:42:26  <delt0r_> antimatroid, that was the orginal idea....
2010-11-07T06:42:38  <delt0r_> wasn't ment to be fun, just good for debug
2010-11-07T06:43:11  <delt0r_> but others liked the idea of a better "select all" type thing
2010-11-07T06:43:15  <delt0r_> Well
2010-11-07T06:43:27  <delt0r_> I can look at adding a specifiy exact numbers
2010-11-07T06:43:36  <antimatroid> hmm fair enough, i could debug stuff a million times better if i could send exact ship numbers
2010-11-07T06:43:49  <antimatroid> but i'm grateful for it's extistence at all :()
2010-11-07T06:43:54  <antimatroid> :)***
2010-11-07T06:43:57  <delt0r_> however i am crazy busy right now and will be till midweek
2010-11-07T06:44:11  <antimatroid> yeah, i have exams the next two days, i'm not doing any work atm anyway :)
2010-11-07T06:44:21  <delt0r_> good news is that the cluster is empty
2010-11-07T06:44:31  <delt0r_> so i have 500 cores to get my work done
2010-11-07T06:44:55  <delt0r_> so i could be finished by monday...
2010-11-07T06:45:09  <delt0r_> will post to the thread anyway once done
2010-11-07T06:46:05  <antimatroid> nice
2010-11-07T06:50:55  <lavalamp_> there's not enough sucky bots on tcp atm :9
2010-11-07T06:50:58  <lavalamp_> er, :(
2010-11-07T06:53:04  <nullkuhl> delt0r_: well i find the +X showing the growth rate really annoying, i was about to remove it from the src code in my local copy of jbotmanager, perhaps a check box to show it/hide it ?
2010-11-07T06:53:25  <delt0r_> ? sorry
2010-11-07T06:53:30  <delt0r_> not following the +X
2010-11-07T06:53:33  <nullkuhl> in the visualization
2010-11-07T06:53:39  <nullkuhl> each planet has a string
2010-11-07T06:53:41  <delt0r_> oh
2010-11-07T06:53:45  <delt0r_> right
2010-11-07T06:53:48  <delt0r_> why?
2010-11-07T06:54:02  <nullkuhl> well the size of the planets tells that already
2010-11-07T06:54:10  <delt0r_> not well IMO
2010-11-07T06:54:52  <nullkuhl> other visualizers doesnt show it as well, sometimes in human mode, i tried focus on my current ships and its not easy to spot them by the eye in the string along with +X there
2010-11-07T06:55:22  <nullkuhl> especially that the string gets out of the planet border most of the time
2010-11-07T06:55:32  <nullkuhl> and enlarging the planet size makes them intersect which is worse.
2010-11-07T06:56:26  <lavalamp_> what I wish jbotmanager showed was planet ids... but I do fancy logging so it's ok: http://pastebin.com/puKMKSrq
2010-11-07T06:56:55  <delt0r_> I will look into --Perhaps its best to post to the thread on the forums... I really wont get to it till later... that way i have a list of requrested fetures
2010-11-07T06:58:04  <sigh> planet ids in a visualiser are invaluable when debugging
2010-11-07T06:58:37  <nullkuhl> lavalamp_: wow, its almost impossibel to get any useful info from a 1 second look at any planet
2010-11-07T06:58:42  <delt0r_> dam
2010-11-07T06:59:00  <delt0r_> My cluster is fast... Perhaps i will get to it this afternoon
2010-11-07T06:59:27  <nullkuhl> lavalamp_: may be at least they can be written in different levels/lines
2010-11-07T06:59:38  <nullkuhl> different font styles/size as well
2010-11-07T06:59:45  <lavalamp_> I toggle on and off different info depending on what I'm debugging... like if I'm debugging front line calculation, it only shows those planets...
2010-11-07T06:59:52  <delt0r_> nullkuhl, probably not going to happen
2010-11-07T06:59:59  <delt0r_> simple toggle on off... yes
2010-11-07T06:59:59  <nullkuhl> delt0r_: whats that ?
2010-11-07T07:00:05  <delt0r_> fancy font crap
2010-11-07T07:00:07  <delt0r_> nope
2010-11-07T07:00:12  <nullkuhl> lol
2010-11-07T07:00:28  <delt0r_> sure you have a point for a good *polished* gui
2010-11-07T07:00:34  <delt0r_> but its just not that
2010-11-07T07:00:59  <nullkuhl> am just talking from a designer point of view, anyway its great that jbot is opensrc, i can do some changes locally and show screenshots to everyone then
2010-11-07T07:01:01  <delt0r_> however feel free to write a patch... the code is a little hacky... but not too bad...
2010-11-07T07:05:24  *** Utkarsh <Utkarsh!~admin@117.201.87.97> has quit IRC (Ping timeout: 264 seconds)
2010-11-07T07:11:40  *** Utkarsh <Utkarsh!~admin@117.201.82.198> has joined #aichallenge
2010-11-07T07:15:39  <contestbot> Top 10 players: bocsimacko(4348), shangas(4200), asavis(4111), Raschi(4105), bix0r4ever(4100), rsergio(4078), george(4072), dmj111(4033), Xrillo(4027), GreenTea(4019)
2010-11-07T07:30:23  *** Bobng <Bobng!~T@78.148.53.228> has joined #aichallenge
2010-11-07T07:32:42  *** Bobng <Bobng!~T@78.148.53.228> has quit IRC (Client Quit)
2010-11-07T07:38:45  *** Xanthor[aw] <Xanthor[aw]!~Xanthor@ALyon-257-1-19-161.w86-209.abo.wanadoo.fr> has quit IRC (Ping timeout: 255 seconds)
2010-11-07T07:39:35  *** Aaaarg <Aaaarg!~Aaaarg@203-100-216-90.callplus.net.nz> has quit IRC (Ping timeout: 240 seconds)
2010-11-07T07:50:36  *** Appleman1234 <Appleman1234!~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au> has quit IRC (Ping timeout: 255 seconds)
2010-11-07T07:55:29  *** Xanthor[aw] <Xanthor[aw]!~Xanthor@ALyon-257-1-55-167.w90-15.abo.wanadoo.fr> has joined #aichallenge
2010-11-07T08:03:32  *** siconize <siconize!55b61664@gateway/web/freenode/ip.85.182.22.100> has joined #aichallenge
2010-11-07T08:11:01  *** irchs <irchs!~Jan@2002:5ec2:21a8:0:226:bbff:fe1d:f630> has quit IRC (Quit: irchs)
2010-11-07T08:12:57  <bhasker> did daylight savings start today?
2010-11-07T08:15:40  <contestbot> Top 10 players: bocsimacko(4357), shangas(4215), bix0r4ever(4115), asavis(4111), Raschi(4090), fglider(4082), rsergio(4078), george(4078), dmj111(4035), Xrillo(4019)
2010-11-07T08:20:22  *** siconize <siconize!55b61664@gateway/web/freenode/ip.85.182.22.100> has quit IRC (Ping timeout: 265 seconds)
2010-11-07T08:42:41  *** hellman <hellman!~hellman@nat232-248-205-109.tvoe.tv> has quit IRC (Remote host closed the connection)
2010-11-07T08:49:36  *** fruit <fruit!5f853906@gateway/web/freenode/ip.95.133.57.6> has joined #aichallenge
2010-11-07T08:56:56  *** fruit <fruit!5f853906@gateway/web/freenode/ip.95.133.57.6> has left #aichallenge
2010-11-07T08:59:56  *** sigh <sigh!~sigh@115.114-50-210.dynamic.dsl.mel.iprimus.net.au> has quit IRC (Remote host closed the connection)
2010-11-07T09:03:54  <antimatroid> daylight savings started a few weeks ago for me
2010-11-07T09:03:59  <antimatroid> wouldn't you just be losing it now?
2010-11-07T09:15:39  <contestbot> Top 10 players: bocsimacko(4371), shangas(4244), bix0r4ever(4123), Raschi(4096), george(4089), fglider(4089), asavis(4082), rsergio(4072), dmj111(4032), Xrillo(4030)
2010-11-07T09:22:08  *** perestrelka <perestrelka!~vlad@194.242.5.47> has quit IRC (Ping timeout: 250 seconds)
2010-11-07T09:24:45  *** perestrelka <perestrelka!~vlad@194.242.5.47> has joined #aichallenge
2010-11-07T09:28:19  *** Bobng <Bobng!~T@78.148.53.228> has joined #aichallenge
2010-11-07T09:35:57  *** siconize <siconize!55b61664@gateway/web/freenode/ip.85.182.22.100> has joined #aichallenge
2010-11-07T09:43:19  *** irchs <irchs!~Jan@2002:5ec2:21a8:0:226:bbff:fe1d:f630> has joined #aichallenge
2010-11-07T09:44:58  *** wh1teside_ <wh1teside_!~admin@adsl-70-131-50-177.dsl.emhril.sbcglobal.net> has joined #aichallenge
2010-11-07T09:48:08  *** wh1teside <wh1teside!~admin@adsl-70-131-52-125.dsl.emhril.sbcglobal.net> has quit IRC (Ping timeout: 252 seconds)
2010-11-07T09:48:33  <nullkuhl> delt0r_: still around ?
2010-11-07T09:48:44  <delt0r_> yep
2010-11-07T09:49:20  <nullkuhl> in which class do you manage writing on the visualised game (i want to control what is written on the planets at visualization)
2010-11-07T09:49:55  <delt0r_> GameViewPanel
2010-11-07T09:50:08  <delt0r_> in the hacky paint component method
2010-11-07T09:50:48  <Prillicy> alright, g'mornin'
2010-11-07T09:51:04  <Prillicy> hi delt0r, thanks for your visualizer
2010-11-07T09:51:12  <delt0r_> np//
2010-11-07T09:51:28  <delt0r_> seems to be my birthday or something
2010-11-07T09:51:34  <delt0r_> everyone is thanking me
2010-11-07T09:51:37  <delt0r_> :D
2010-11-07T09:53:39  <nullkuhl> delt0r_: in your birthday, you're usually thanking everyone for their gifts :P
2010-11-07T09:53:40  <Prillicy> flat question to anyone here, I seem to be having trouble turning the lists that the game spits out into arrays so I can loop through them
2010-11-07T09:54:42  <nullkuhl> allright what do you think http://i54.tinypic.com/jzyyir.jpg ?
2010-11-07T09:57:11  <delt0r_> nullkuhl, Only the growth has been "turned" off... i don't see where i turn them back on?
2010-11-07T09:58:34  <nullkuhl> yes i didnt add that yet, it was just a 1 minute playing with the colors nothing more
2010-11-07T10:02:52  <delt0r_> right
2010-11-07T10:03:18  <nullkuhl> bhasker, antimatroid, delt0r_, Prillicy: any one liks the new coloring ? http://i54.tinypic.com/jzyyir.jpg
2010-11-07T10:03:27  <delt0r_> well i don't like black backgrounds --but each to his own...
2010-11-07T10:04:21  <nullkuhl> delt0r_: aha, well almost all visualizers have dark bgs anyway, its more easy for the eyes, i set my irc client bg to black as well here.
2010-11-07T10:05:11  <delt0r_> I find its worse... you also can't read things easily... there is a reason why paper is white...
2010-11-07T10:06:50  <nullkuhl> paper is white because its costy to color them black :), and its even harder to write in white
2010-11-07T10:07:07  <Prillicy> nice
2010-11-07T10:07:09  <delt0r_> in fact its not... white ink is cheaper...
2010-11-07T10:07:18  <delt0r_> making paper white is expensive
2010-11-07T10:07:33  <delt0r_> but as i said
2010-11-07T10:07:38  <delt0r_> each to his own
2010-11-07T10:08:04  <nullkuhl> yeah, its just that ui design generally black bg is more comforting for the eyes.
2010-11-07T10:08:12  <nullkuhl> as black doesnt reflect any light towards the eyes
2010-11-07T10:08:19  <delt0r_> but please don't be one of the people that give presentations with black slides with white text ;) --most cannot read them :P
2010-11-07T10:08:47  <delt0r_> nullkuhl, word, XL and just about all apps i have seen do not use blk bg
2010-11-07T10:09:06  <delt0r_> in fact --I can't think of any that do
2010-11-07T10:09:53  <nullkuhl> yes because few people have sight problems and  white text on black bg is unseenable to them, however for the rest its quite the opposite
2010-11-07T10:10:07  <nullkuhl> all NVidia's latest slides are white text on black bg for instance
2010-11-07T10:10:10  <delt0r_> thats bs...
2010-11-07T10:10:28  <delt0r_> any--this is ot...
2010-11-07T10:12:19  <nullkuhl> delt0r_: http://xpt.sourceforge.net/techdocs/misc/ce01-DarkBackgroundIsGoodForYou/  :P
2010-11-07T10:12:56  <delt0r_> Thats hardy data... is opinion.
2010-11-07T10:12:58  *** amriedle <amriedle!~amriedle@CPE0013f7f2be12-CM0013f7f2be0e.cpe.net.cable.rogers.com> has joined #aichallenge
2010-11-07T10:13:33  <delt0r_> 99.9% of apps don't use black background
2010-11-07T10:13:51  <delt0r_> its not a consprisy by 1% of people with eye problems
2010-11-07T10:14:54  <nullkuhl> well it's just that majority doesn't necessarily mean right, its more of a medical issue here related to light reflection and sight
2010-11-07T10:15:26  <delt0r_> well there is no medical reason or data to sujest one is better than the other
2010-11-07T10:15:33  <delt0r_> guess where i work ;)
2010-11-07T10:15:41  <contestbot> Top 10 players: bocsimacko(4367), shangas(4240), fglider(4155), bix0r4ever(4127), asavis(4085), Raschi(4081), rsergio(4080), george(4074), Xrillo(4030), GreenTea(4018)
2010-11-07T10:15:45  <nullkuhl> where ?
2010-11-07T10:15:50  <delt0r_> lol
2010-11-07T10:16:07  <nullkuhl> just wondering seriously
2010-11-07T10:16:17  <raedwulf> i like light text on dark background
2010-11-07T10:16:22  <raedwulf> strains my eyes less
2010-11-07T10:16:27  <delt0r_> I work as a Popgenetics with the math department and the medical school
2010-11-07T10:16:31  <raedwulf> i invert my pdfs time to time
2010-11-07T10:16:55  <delt0r_> I often do stats work for medical studies
2010-11-07T10:17:08  <nullkuhl> the very same here raedwulf, far from this topic, do you like the new coloring fo delt0r_'s visualizer ?
2010-11-07T10:17:10  <delt0r_> most medical studies are total crap
2010-11-07T10:17:17  <raedwulf> no idea
2010-11-07T10:17:20  <raedwulf> where is it
2010-11-07T10:17:29  * raedwulf scrolls up[
2010-11-07T10:17:35  <nullkuhl> [16:54] <nullkuhl> allright what do you think http://i54.tinypic.com/jzyyir.jpg ?
2010-11-07T10:17:35  <raedwulf> yea its nice
2010-11-07T10:18:59  <raedwulf> i really want to get round with hacking a web browser up to invert colours on certain webpages
2010-11-07T10:19:07  *** irchs <irchs!~Jan@2002:5ec2:21a8:0:226:bbff:fe1d:f630> has quit IRC (Quit: irchs)
2010-11-07T10:19:37  <raedwulf> without breaking the website, i.e. just a simple filter over the completely rasterised site
2010-11-07T10:20:02  <nullkuhl> raedwulf: there is a workaround for this anyway,  when i work on ubuntu i activate the beryl fx that inverts the whole screen
2010-11-07T10:20:27  <nullkuhl> this way i get all my docs, websites inverted for me, and remains the same for others
2010-11-07T10:20:51  <antimatroid> that question was rank
2010-11-07T10:21:11  <nullkuhl> what ?
2010-11-07T10:22:34  <antimatroid> http://imgur.com/fmQ4L.png
2010-11-07T10:22:40  <raedwulf> nullkuhl: nah, i don't usually want *everything* inverted
2010-11-07T10:22:48  <antimatroid> using AGM inequality
2010-11-07T10:22:59  <antimatroid> it was pretty easy once i worked out how it's done, but it took me a while
2010-11-07T10:23:52  <delt0r_> not induction antimatroid ?
2010-11-07T10:23:58  <antimatroid> nope
2010-11-07T10:24:13  <delt0r_> that would have been my first atempt
2010-11-07T10:24:48  <nullkuhl> raedwulf: you can pick the windows that you want to invert specifically without having to invert the whole screen ;)
2010-11-07T10:25:35  <antimatroid> you have prod rt(i(n-i)) = prod rt(i)^2, then by AGM inequality, you get sum LHS <= prod rt(i )^2 = sum rt(i )^2 and that gives it
2010-11-07T10:25:57  <antimatroid> oh, and n(n-1)/2 is just sum 1 to n of i
2010-11-07T10:26:02  <antimatroid> normal trinagle number thing
2010-11-07T10:26:45  <antimatroid> like i said, gross, but easy once one got there
2010-11-07T10:26:58  <delt0r_> it always is
2010-11-07T10:27:07  <antimatroid> if only
2010-11-07T10:27:11  <delt0r_> Maths is only hard when you don't get it
2010-11-07T10:27:30  <delt0r_> then its easy and its hard to see how it was hard
2010-11-07T10:27:45  <antimatroid> i often understand stuff but find it hard to remember it all when under a test condition
2010-11-07T10:27:48  <antimatroid> i suck at exams
2010-11-07T10:27:52  <delt0r_> which in turn makes it hard to explain
2010-11-07T10:28:04  <delt0r_> I do ok at exams.
2010-11-07T10:28:07  <delt0r_> did
2010-11-07T10:28:09  <antimatroid> especially my analysis exam the other day, there was just lots to remember
2010-11-07T10:28:15  <delt0r_> I don't do them anymore
2010-11-07T10:28:31  <delt0r_> analysis is hard ;)
2010-11-07T10:29:00  <antimatroid> yeah, algebra and analysis ended up being my two hardest undergrad units
2010-11-07T10:29:13  <antimatroid> i've got a pretty rough line up next year though
2010-11-07T10:29:41  <antimatroid> ring theory, abelian groups and modules, functional analysis, topology, symmetry principles and group theory in physics, and dynamical systems
2010-11-07T10:29:51  <antimatroid> i sat through a measure theory class this semester too :| head asplode
2010-11-07T10:31:09  <delt0r_> sounds like fun
2010-11-07T10:31:37  <delt0r_> I did Galios fields for my masters.
2010-11-07T10:32:19  <delt0r_> I only lernt enough measure theory to know whats a "no no"
2010-11-07T10:40:01  *** Frontier <Frontier!~Marco_Lei@dslb-088-071-242-074.pools.arcor-ip.net> has quit IRC (Read error: Connection reset by peer)
2010-11-07T10:40:10  *** Frontier <Frontier!~Marco_Lei@dslb-088-071-242-074.pools.arcor-ip.net> has joined #aichallenge
2010-11-07T10:54:34  <amriedle> The real part of the AI challenge is to determine the question, can blind people see that taste of cinnamon toast crunch... so far inconclusive... lol
2010-11-07T10:54:40  *** irchs <irchs!~Jan@2002:5ec2:21a8:0:226:bbff:fe1d:f630> has joined #aichallenge
2010-11-07T10:56:21  *** boegel <boegel!~boegel@vpnq243.ugent.be> has quit IRC (Quit: This computer has gone to sleep)
2010-11-07T11:04:01  <antimatroid> amriedle: wtf?
2010-11-07T11:04:33  <antimatroid> your name is an anagaram for im a ridle
2010-11-07T11:04:48  <antimatroid> ridlee*
2010-11-07T11:04:49  <antimatroid> doh
2010-11-07T11:04:55  <nullkuhl> riddle*
2010-11-07T11:05:05  <antimatroid> there aren't two d's :P
2010-11-07T11:05:26  <nullkuhl> there is
2010-11-07T11:05:31  <antimatroid> not in his username
2010-11-07T11:05:43  <nullkuhl> haha
2010-11-07T11:10:41  <amriedle> you may want to type "can blind people see" on google and see the suggestions
2010-11-07T11:11:19  <amriedle> you may want to not insult my last name, i might ... do nothing
2010-11-07T11:12:02  *** Commandir <Commandir!~joakim@h85-8-40-32.static.se.alltele.net> has left #aichallenge
2010-11-07T11:12:13  <amriedle> you may want to think that i should start all sentences with "you may want to"
2010-11-07T11:12:25  *** siconize <siconize!55b61664@gateway/web/freenode/ip.85.182.22.100> has quit IRC (Ping timeout: 265 seconds)
2010-11-07T11:12:33  <raedwulf> you may want to stop
2010-11-07T11:12:48  <amriedle> you may want to make me stop, but i may not want to.
2010-11-07T11:13:57  <amriedle> btw my initial are AMR <-- hence amriedler
2010-11-07T11:14:06  * raedwulf hands amriedle a hand-grenade and says, "you may want to pull the pin on this."
2010-11-07T11:14:29  <amriedle> truncated of course
2010-11-07T11:14:42  <raedwulf> you're an idler?
2010-11-07T11:15:03  <amriedle> or a riddler :P, depends on the typo
2010-11-07T11:15:12  <raedwulf> hehe
2010-11-07T11:15:41  <contestbot> Top 10 players: bocsimacko(4379), shangas(4269), bix0r4ever(4161), asavis(4151), rsergio(4120), george(4072), fglider(4064), Xrillo(4035), dmj111(4030), GreenTea(4002)
2010-11-07T11:15:57  <raedwulf> delimare? realdime?
2010-11-07T11:17:05  *** Rubicon-|-Cross <Rubicon-|-Cross!~RubiconCr@adsl-99-30-224-80.dsl.pltn13.sbcglobal.net> has quit IRC (Quit: Leaving)
2010-11-07T11:17:27  *** irchs <irchs!~Jan@2002:5ec2:21a8:0:226:bbff:fe1d:f630> has quit IRC (Quit: irchs)
2010-11-07T11:18:27  <raedwulf> readmile
2010-11-07T11:18:40  *** irchs <irchs!~Jan@2002:5ec2:21a8:0:226:bbff:fe1d:f630> has joined #aichallenge
2010-11-07T11:19:58  *** irchs <irchs!~Jan@2002:5ec2:21a8:0:226:bbff:fe1d:f630> has quit IRC (Client Quit)
2010-11-07T11:20:40  *** Utkarsh <Utkarsh!~admin@117.201.82.198> has quit IRC (Ping timeout: 265 seconds)
2010-11-07T11:22:28  *** tapwater <tapwater!~tapwater@cpe-70-112-195-55.austin.res.rr.com> has joined #aichallenge
2010-11-07T11:28:43  *** Utkarsh <Utkarsh!~admin@117.201.89.212> has joined #aichallenge
2010-11-07T11:29:07  <Prillicy> bah, I cann't seem to get this bot to get through a game without crashing, and I don't really have the ability to debug because I don't know java :(
2010-11-07T11:30:29  <bhasker> weird a hybrid approach to calculating free ships works better
2010-11-07T11:30:34  <bhasker> or at least seems to
2010-11-07T11:31:24  <bhasker> a strict account for everything is too conservative and starts doing nothing once you start losing
2010-11-07T11:38:04  *** Utkarsh <Utkarsh!~admin@117.201.89.212> has quit IRC (Ping timeout: 265 seconds)
2010-11-07T11:43:35  *** Utkarsh <Utkarsh!~admin@117.201.80.244> has joined #aichallenge
2010-11-07T11:48:45  *** Utkarsh <Utkarsh!~admin@117.201.80.244> has quit IRC (Ping timeout: 276 seconds)
2010-11-07T11:50:30  *** irchs <irchs!~Jan@94-194-33-168.zone8.bethere.co.uk> has joined #aichallenge
2010-11-07T11:51:19  *** demerzel_ <demerzel_!488039c1@gateway/web/freenode/ip.72.128.57.193> has joined #aichallenge
2010-11-07T11:51:37  <demerzel_> woop i'm where everyone else was a month ago
2010-11-07T11:51:59  <demerzel_> gonna need a rocky montage
2010-11-07T11:52:39  <demerzel_> i hate daylight savings
2010-11-07T11:52:58  <demerzel_> it saves it for me to look at outside my work's window till march :(
2010-11-07T11:54:35  *** Utkarsh <Utkarsh!~admin@117.201.82.103> has joined #aichallenge
2010-11-07T11:55:59  *** perror <perror!~fleury@aldebaran.labri.fr> has joined #aichallenge
2010-11-07T11:56:10  <lavalamp_> humph, it's not that my bot *can't* defend itself, it's that it did the "math" and decided it wsan't worth it, grrr
2010-11-07T11:56:50  *** bhasker_ <bhasker_!~bhasker@adsl-76-254-15-251.dsl.pltn13.sbcglobal.net> has joined #aichallenge
2010-11-07T11:56:50  *** bhasker <bhasker!~bhasker@adsl-76-254-15-251.dsl.pltn13.sbcglobal.net> has quit IRC (Read error: Connection reset by peer)
2010-11-07T12:01:05  <amriedle> heheh
2010-11-07T12:02:47  *** phreeza <phreeza!~phreeza@f054211196.adsl.alicedsl.de> has joined #aichallenge
2010-11-07T12:03:01  <raedwulf> bhasker_: thanks!
2010-11-07T12:03:12  <raedwulf> i was wondering how that approach would work
2010-11-07T12:04:39  <demerzel_> well he ran into the same issue i did: if you absolutely know that anywhere you go the enemy can put more ships there and that's a criteria you lock down
2010-11-07T12:04:49  <demerzel_> if you ever lose the advantage that is
2010-11-07T12:05:35  <bhasker_> yup
2010-11-07T12:06:04  <raedwulf> i suppose if you lose advantage tho you technically lose if you play an optimimum opponent
2010-11-07T12:06:11  <raedwulf> optimum*
2010-11-07T12:06:13  <raedwulf> lol
2010-11-07T12:06:22  <demerzel_> yea that's true
2010-11-07T12:06:52  <demerzel_> but you can make your bot perfect but if the other bot isn't and somehow gains an advantage because of it
2010-11-07T12:06:58  <demerzel_> if you can't account for that then you're screwed
2010-11-07T12:07:12  <demerzel_> whereas you might not be otherwise
2010-11-07T12:07:29  <raedwulf> thats true
2010-11-07T12:07:38  <raedwulf> hopefully my technique will handle this
2010-11-07T12:07:40  <demerzel_> judging by a lot of the eval stuff i've seen they're playing with relative numbers that have no bearing on the actual game
2010-11-07T12:07:49  <demerzel_> beyond a 'good/bad' eval
2010-11-07T12:07:54  <raedwulf> but hell, its not easy using the code i've written xD
2010-11-07T12:08:02  *** irchs <irchs!~Jan@94-194-33-168.zone8.bethere.co.uk> has quit IRC (Quit: irchs)
2010-11-07T12:08:03  <demerzel_> so you're a shitty coder?
2010-11-07T12:08:04  <demerzel_> that sucks
2010-11-07T12:08:12  <raedwulf> no, its not that, its because it complicates things
2010-11-07T12:08:20  <demerzel_> lol
2010-11-07T12:08:34  <raedwulf> but its the only way i could think of doing it (at the time)
2010-11-07T12:08:44  <raedwulf> i have an idea to optimise it, might even simplify it
2010-11-07T12:08:53  <raedwulf> but probably way too much for too little gain
2010-11-07T12:09:55  <demerzel_> what is 'it'
2010-11-07T12:10:01  <demerzel_> i'm in C it's already optimized lol
2010-11-07T12:10:16  <demerzel_> that's the lovely thing about C the lowest effort way is usually the most optimized
2010-11-07T12:10:23  <raedwulf> hehe ^^
2010-11-07T12:15:42  <contestbot> Top 10 players: bocsimacko(4378), shangas(4272), bix0r4ever(4175), asavis(4159), rsergio(4123), fglider(4072), george(4063), dmj111(4043), Xrillo(4034), Accoun(4015)
2010-11-07T12:17:35  <amriedle> i do not agree with such statement :)
2010-11-07T12:17:54  <delt0r_> neither do i
2010-11-07T12:20:29  *** hellman <hellman!~hellman@nat232-248-205-109.tvoe.tv> has joined #aichallenge
2010-11-07T12:22:40  <demerzel_> what's it like being wrong
2010-11-07T12:25:26  <demerzel_> hmm i am debating using malloc/free vs just declaring a big array of nodes
2010-11-07T12:25:51  <demerzel_> seems like the latter would be more optimal but i guess then i gotta have some func to make the array bigger
2010-11-07T12:26:53  <phreeza> demerzel_: are you using straight up C?
2010-11-07T12:27:01  <demerzel_> yup iso-c
2010-11-07T12:27:33  <demerzel_> i would use ansi-c but vs doesn't support it plus gcc has its own implementation of c99 that i am not familiar with
2010-11-07T12:27:51  <demerzel_> not that many great differences between c89 and c99 ;)
2010-11-07T12:28:04  <demerzel_> i am toying with the idea of a pointer / realloc to make a dynamic array
2010-11-07T12:28:08  <phreeza> ah ok :)
2010-11-07T12:28:11  *** davidjliu <davidjliu!473a6d0e@gateway/web/freenode/ip.71.58.109.14> has joined #aichallenge
2010-11-07T12:28:18  <phreeza> because i saw a nifty thing a while back
2010-11-07T12:28:27  <phreeza> an allocator for stl containers
2010-11-07T12:28:36  <phreeza> that puts as much as you like on the stack
2010-11-07T12:28:59  <phreeza> and then if you exceepd that, puts the resto on the heap automatically
2010-11-07T12:29:24  <demerzel_> that could be implemented relatively easily in C
2010-11-07T12:29:35  <demerzel_> you declare a static array on the stack then use malloc beyond those dims
2010-11-07T12:29:51  <phreeza> yea
2010-11-07T12:29:55  <demerzel_> i was thinking about that too
2010-11-07T12:30:03  *** nann <nann!~apophis@80-219-211-230.dclient.hispeed.ch> has joined #aichallenge
2010-11-07T12:30:06  <phreeza> probably not that hard, youre right
2010-11-07T12:30:09  <demerzel_> have a [500] node array on the stack then handle 500+
2010-11-07T12:30:47  <demerzel_> i usually avoid malloc like the plague
2010-11-07T12:30:47  <phreeza> you'd need a getter function i guess
2010-11-07T12:30:52  <demerzel_> but i think i'm gonna have to mess with it
2010-11-07T12:31:08  <phreeza> heh
2010-11-07T12:31:15  <demerzel_> phreeza: yea essentiallly you'd wrap node[x] to be get_node(x) and handle it that way
2010-11-07T12:31:17  <phreeza> for performance tuning?
2010-11-07T12:31:29  <demerzel_> no because i can't know how many nodes i will be looking at
2010-11-07T12:31:57  <demerzel_> of course the entire memory footprint of my bot atm is <1k maybe i'll just declare a 12! array of nodes haha
2010-11-07T12:32:02  <phreeza> ah, right...
2010-11-07T12:32:10  <phreeza> hehhehe
2010-11-07T12:32:33  *** Apophis <Apophis!~apophis@80-219-211-230.dclient.hispeed.ch> has quit IRC (Ping timeout: 272 seconds)
2010-11-07T12:32:59  <demerzel_> oh <10k not 1k hehe
2010-11-07T12:33:08  <demerzel_> still - damn small
2010-11-07T12:33:30  <phreeza> yes thats really quite small
2010-11-07T12:34:35  <demerzel_> well i am just now getting into trees etc
2010-11-07T12:34:43  <demerzel_> i am sure it will rise exponentially lol
2010-11-07T12:35:08  <phreeza> yea my tree bot gets up to 100M easily i think
2010-11-07T12:35:18  <phreeza> its very inefficient though
2010-11-07T12:36:01  <phreeza> but i think its faster to just keep a lot of the stuff hanging around instead of freeing it sometimes though
2010-11-07T12:37:34  <demerzel_> damn 100M?  how big are your nodes
2010-11-07T12:37:34  *** yasith <yasith!~yasith@112.135.109.15> has quit IRC (Ping timeout: 240 seconds)
2010-11-07T12:38:38  <demerzel_> my entire gamestate is ~13k
2010-11-07T12:39:21  <phreeza> i'm not sure but each node contains a complete state of the game more or less :D
2010-11-07T12:39:22  <demerzel_> i can compress it down to ~2k since i'm not looking at all 201 turns but haven't needed to yet
2010-11-07T12:39:54  <demerzel_> ah i spent a good amount of time abstracting the static game stuff from the dynamic
2010-11-07T12:40:13  <demerzel_> my gamestate consists of ships, supply, and incoming - all 23xint[201]
2010-11-07T12:40:41  <demerzel_> supply is a logical one tho but it's useful for evaluating whether i can do stuff later on
2010-11-07T12:40:59  <phreeza> yea i basically took the Game object from alberts C++ lib and used that
2010-11-07T12:41:38  <demerzel_> link?
2010-11-07T12:41:45  <phreeza> just being lazy there is probalbly a lot of room for improvement
2010-11-07T12:42:14  <phreeza> https://github.com/albertz/planet_wars-cpp
2010-11-07T12:43:01  <demerzel_> oh you are using c++
2010-11-07T12:43:26  <phreeza> yea
2010-11-07T12:43:41  <phreeza> and i have about 15k nodes
2010-11-07T12:43:45  <demerzel_> aiee yea that can use a lot of improvement
2010-11-07T12:44:01  <demerzel_> vectors of structs
2010-11-07T12:44:05  <phreeza> the memory consumption is about 25M i just checked
2010-11-07T12:44:13  <demerzel_> you have no performance issues with 15k nodes of that?
2010-11-07T12:44:17  <phreeza> so its about 1k per node i guess
2010-11-07T12:44:19  *** irchs <irchs!~Jan@2002:5ec2:21a8:0:226:bbff:fe1d:f630> has joined #aichallenge
2010-11-07T12:44:21  <demerzel_> damn i can probably go futher than i thought
2010-11-07T12:44:54  <demerzel_> are you storing gamestate in the node?
2010-11-07T12:44:55  <nullkuhl> wow, i never thought that its even possible to reach some finite tree that is small enough to work in this game
2010-11-07T12:45:23  <demerzel_> nullkuhl: it takes quite a bit of effort to get to that point lol
2010-11-07T12:45:37  <phreeza> nullkuhl: it isn't particularly good but it does win some games, which is really all i wanted to prove with the bot ;)
2010-11-07T12:46:05  <nullkuhl> the nodes are game states ? like map of planets + their ships etc.
2010-11-07T12:46:09  <phreeza> http://72.44.46.68/canvas?game_id=897759
2010-11-07T12:46:13  <phreeza> yep
2010-11-07T12:46:16  <demerzel_> what is a planet in that implementation?  just 5 ints?
2010-11-07T12:46:20  <demerzel_> i guess 1k might be true
2010-11-07T12:46:41  <demerzel_> my gamestate includes all population progressions
2010-11-07T12:47:27  <phreeza> the expensive part is not the storing etc though, what costs me is that to assign a node its value i play it until all fleets have arrived
2010-11-07T12:47:31  <phreeza> with a for loop
2010-11-07T12:47:44  <phreeza> my guess is thats where i could really make some improvements
2010-11-07T12:48:04  <phreeza> by skipping ahead moves or something like that
2010-11-07T12:48:23  <nullkuhl> i wonder how is the tree finite anyway, do you account for any possible moves made by enemy player ? for example sending 1 to N ships from any planet ?
2010-11-07T12:48:51  <phreeza> i generate a certain number of moves drawn from a random distribution
2010-11-07T12:49:15  <phreeza> its by no means a complete search
2010-11-07T12:49:19  <phreeza> monte carly
2010-11-07T12:49:23  <phreeza> carlo
2010-11-07T12:49:24  <phreeza> lol
2010-11-07T12:50:12  *** yasith <yasith!~yasith@112.135.73.41> has joined #aichallenge
2010-11-07T12:50:31  <nullkuhl> phreeza: whats your rank in official server ?
2010-11-07T12:50:44  <phreeza> its not running on official
2010-11-07T12:51:11  <phreeza> i have a heuristic bot on there which i havent touched in >1 month
2010-11-07T12:51:20  <phreeza> toying with some other ideas right now
2010-11-07T12:52:02  <nullkuhl> aha
2010-11-07T12:52:28  <phreeza> but not really getting anything done
2010-11-07T12:52:48  <demerzel_> phreeza: my bot on official is C# and like 2 months old lol
2010-11-07T12:53:04  <phreeza> heh mine is python
2010-11-07T12:53:37  <phreeza> its just an adapted starter pack really, but it did remarably well for that in the first week or so ;)
2010-11-07T12:54:04  <phreeza> then people figured out staging and sniping :) and mine doesn't do either of those
2010-11-07T12:58:29  <phreeza> demerzel_: so the main problem i see with this tree search is that it can make good tactical decisions, because you can probably search the next 3-4 moves to a reasonable degree. but what it doesn't know how to do well is plan for the future, because the decison tree cant be built for say 20 moves into the future with any breadth. do you have any idea how you want to tackle that?
2010-11-07T12:58:54  *** RainCT <RainCT!~RainCT@ubuntu/member/pdpc.student.rainct> has joined #aichallenge
2010-11-07T13:01:22  <demerzel_> phreeza: haha not there yet
2010-11-07T13:01:38  <phreeza> :)
2010-11-07T13:02:21  <Error323> good evening
2010-11-07T13:02:33  <phreeza> hi
2010-11-07T13:02:42  <Error323> how are things going
2010-11-07T13:03:01  <phreeza> meh..
2010-11-07T13:04:28  <demerzel_> building a b-tree in C
2010-11-07T13:04:36  <demerzel_> not too bad!
2010-11-07T13:08:08  *** Appleman1234 <Appleman1234!~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au> has joined #aichallenge
2010-11-07T13:09:13  *** TravisD <TravisD!~Adium@uwsclient-159-148.uws.ualberta.ca> has joined #aichallenge
2010-11-07T13:15:01  *** Sylph <Sylph!~x@adsl-99-121-237-189.dsl.sfldmi.sbcglobal.net> has quit IRC (Read error: Connection reset by peer)
2010-11-07T13:15:42  <contestbot> Top 10 players: bocsimacko(4364), shangas(4261), bix0r4ever(4181), rsergio(4124), fglider(4085), george(4068), dmj111(4059), asavis(4056), GreenTea(4046), Xrillo(4037)
2010-11-07T13:25:07  <demerzel_> hmm is the difference only semantic between (*node).state and node->state
2010-11-07T13:25:11  <demerzel_> for assignment that is
2010-11-07T13:25:36  <demerzel_> for some reason the former works but the latter doesn't lol
2010-11-07T13:30:26  <raedwulf> demerzel_: thats weird
2010-11-07T13:31:19  <demerzel_> probably a vc quirk
2010-11-07T13:31:23  <demerzel_> NULL is 0 too :(
2010-11-07T13:31:33  <raedwulf> well its supposed to be  :P
2010-11-07T13:31:48  <demerzel_> NULL should be (void*(0)) in C
2010-11-07T13:31:49  <raedwulf> demerzel_: could you examine the difference in asm?
2010-11-07T13:31:59  <raedwulf> oh right, yes that is true
2010-11-07T13:32:06  <raedwulf> you mean its defined simply as 0?
2010-11-07T13:32:09  <demerzel_> nah it doesn't give me that level
2010-11-07T13:32:26  <demerzel_> raedwulf: yes because c_plusplus is defined even though it's not used
2010-11-07T13:32:42  <raedwulf> ah
2010-11-07T13:32:44  <demerzel_> i defined C_NULL manually but then the compiler threw an error - just vc workarounds
2010-11-07T13:33:13  <raedwulf> hehe
2010-11-07T13:42:43  <demerzel_> woo this tree stuff make me feel like such a smart guy!
2010-11-07T13:51:09  <demerzel_> except when it's not working
2010-11-07T13:57:02  *** bhasker_ <bhasker_!~bhasker@adsl-76-254-15-251.dsl.pltn13.sbcglobal.net> has quit IRC (Quit: bhasker_)
2010-11-07T14:02:18  *** Utkarsh <Utkarsh!~admin@117.201.82.103> has quit IRC (Ping timeout: 240 seconds)
2010-11-07T14:07:01  *** Utkarsh <Utkarsh!~admin@117.201.80.110> has joined #aichallenge
2010-11-07T14:07:32  *** Frontier <Frontier!~Marco_Lei@dslb-088-071-242-074.pools.arcor-ip.net> has quit IRC (Ping timeout: 272 seconds)
2010-11-07T14:11:01  *** Mathnerd314 <Mathnerd314!~mathnerd3@dsl.6.28.53.206.cos.dyn.pcisys.net> has quit IRC (Ping timeout: 240 seconds)
2010-11-07T14:15:43  <contestbot> Top 10 players: bocsimacko(4341), shangas(4229), bix0r4ever(4179), rsergio(4095), george(4061), fglider(4036), asavis(4034), dmj111(4034), GreenTea(4032), Xrillo(4022)
2010-11-07T14:16:01  <TravisD> demerzel_: What are you doing with a tree?
2010-11-07T14:24:08  <Migi32> logging
2010-11-07T14:24:12  <Migi32> ba dumm tssh
2010-11-07T14:27:47  *** bhasker <bhasker!~bhasker@64.9.243.180> has joined #aichallenge
2010-11-07T14:28:15  <TravisD> lol
2010-11-07T14:30:11  <bhasker> how goes it Error323
2010-11-07T14:31:02  <Error323> bhasker: hey, well good, i'm working on a new idea
2010-11-07T14:31:17  <Error323> still in a sort of infancy
2010-11-07T14:31:18  <bhasker> your 0.17 is doing quite well
2010-11-07T14:31:26  <bhasker> my new ideas didnt pan out so well lol
2010-11-07T14:31:28  <bhasker> :-\
2010-11-07T14:31:30  <Error323> but preliminaries look good
2010-11-07T14:31:43  <Error323> yeah, its going quite nicely 21 @ official :D
2010-11-07T14:31:50  <Error323> mmm pitty
2010-11-07T14:32:02  <bhasker> i tried to give my bot more accurate information
2010-11-07T14:32:12  <bhasker> as it happens that paralyzes my bot
2010-11-07T14:32:16  <bhasker> if it starts to lose
2010-11-07T14:32:49  <antimatroid> bhasker: i've had the cripple problems before
2010-11-07T14:32:55  <antimatroid> oh god, exam in under 3 hours
2010-11-07T14:32:58  <antimatroid> uni time
2010-11-07T14:33:01  <bhasker> all the best
2010-11-07T14:33:12  <bhasker> yea so i purposely now don't give it accurate information
2010-11-07T14:33:17  <bhasker> so that it does something
2010-11-07T14:33:19  <Error323> bhasker: i had the same problem
2010-11-07T14:33:31  <antimatroid> bhasker: you're doing it wrong
2010-11-07T14:33:40  <antimatroid> you need to make better decisions with accurate information
2010-11-07T14:33:54  <bhasker> well the problem is this
2010-11-07T14:33:59  <bhasker> the way i calculate available ships on a planet
2010-11-07T14:34:13  <Error323> bhasker: try to figure out why its doing worse with your better predictions... _know_ that your predictions themselves are good
2010-11-07T14:34:15  <bhasker> is i try to send all ships of mine + enemies at the planet and see who holds it with how many ships
2010-11-07T14:34:32  <bhasker> for ships in flight i create fleets further out in time
2010-11-07T14:34:35  <bhasker> and resolve all of them
2010-11-07T14:34:48  <bhasker> sadly that factors in growth of neutrals/planets that i may counter in that move
2010-11-07T14:35:05  <bhasker> so if enemy actually started attacking a new neutral or my planet
2010-11-07T14:35:08  <bhasker> i would assume i am going to lose it
2010-11-07T14:35:19  <bhasker> and my availability would goto 0-
2010-11-07T14:35:21  <bhasker> :-\
2010-11-07T14:35:36  <bhasker> even though if i had supported it i would have probably held it
2010-11-07T14:35:42  <antimatroid> what i do it take a dest planet, pick a time i want to take it, implement that then assume we battle it out for the planet at that point, if i am expected to retain it into the future, i return the whole lot as a posisble move
2010-11-07T14:36:09  <bhasker> hmm
2010-11-07T14:36:18  <bhasker> probably i need to stop building an availability list altogether
2010-11-07T14:36:25  <bhasker> i think thats the main problem
2010-11-07T14:36:33  <bhasker> at start of the turn i build like a list of ships i can send from each planet
2010-11-07T14:36:50  <antimatroid> i do that, but also what i can send at each time point
2010-11-07T14:36:58  <bhasker> hmm
2010-11-07T14:37:04  <bhasker> i did that with this new version
2010-11-07T14:37:07  <bhasker> i do it for both me and enemy
2010-11-07T14:37:12  <bhasker> and make them into phantom fleets
2010-11-07T14:37:19  <bhasker> and use them in the battle
2010-11-07T14:37:28  <bhasker> but if the enemy has higher growth
2010-11-07T14:37:37  <bhasker> he will have more future growth/future ships
2010-11-07T14:37:41  <bhasker> which will tip all battles in his favour
2010-11-07T14:37:52  <bhasker> even though currently you may be even on ships
2010-11-07T14:37:56  <Migi32> if the enemy has more growth, 0 ships are available? I have the same problem.
2010-11-07T14:38:03  <bhasker> yes
2010-11-07T14:38:04  <lavalamp_> maybe you need to detect that and go into crazy mode
2010-11-07T14:38:24  <bhasker> so what i do is a hybrid approach now
2010-11-07T14:38:28  <bhasker> when determining available ships
2010-11-07T14:38:32  <bhasker> i only consider the ships on the planets
2010-11-07T14:38:40  <bhasker> and not ships in flight to neutrals
2010-11-07T14:38:41  <Migi32> the trick is of course to prevent the enemy from getting more growth in the first place ;)
2010-11-07T14:38:45  <bhasker> or ships to enemy planet
2010-11-07T14:38:54  <bhasker> hard to do
2010-11-07T14:38:57  <bhasker> sometimes at turn 1
2010-11-07T14:39:01  <bhasker> you just pick bad planets
2010-11-07T14:39:04  <bhasker> then your bot is dead
2010-11-07T14:39:18  <bhasker> cause it will assume that i ca't send anything without losing a planet
2010-11-07T14:39:33  <bhasker> and won't even defend planets
2010-11-07T14:39:38  <bhasker> since there are no available ships
2010-11-07T14:39:49  <Error323> i tried to make better strength predictions by radial falloff of fleet size past a certain amount of time
2010-11-07T14:39:51  <bhasker> so enemy will leisurely pick one planet at a time
2010-11-07T14:40:07  <bhasker> assuming i guess that enemy can't sustain the peak size
2010-11-07T14:46:21  *** Azrathud <Azrathud!~azrathud@c-67-166-43-227.hsd1.co.comcast.net> has joined #aichallenge
2010-11-07T14:54:37  <bhasker> an interesting game to watch
2010-11-07T14:54:39  <bhasker> http://72.44.46.68/canvas?game_id=899073
2010-11-07T14:54:41  <bhasker> though i timed out in the end
2010-11-07T14:54:45  <bhasker> i wonder who would have won
2010-11-07T14:54:51  *** Accoun <Accoun!~1@static-213-88-2-78.chebnet.ru> has quit IRC ()
2010-11-07T14:54:54  <bhasker> looks like i  might have lost though
2010-11-07T14:56:09  <bhasker> we were both chasing each other's tail
2010-11-07T14:56:25  <irchs> :o
2010-11-07T14:56:42  <irchs> hightime has some nice coordinated attack code
2010-11-07T14:56:51  <irchs> Although it doesn't quite work with a map like that
2010-11-07T14:57:25  <bhasker> yea
2010-11-07T15:01:06  *** bhasker_ <bhasker_!~bhasker@69-170-40-2.static-ip.telepacific.net> has joined #aichallenge
2010-11-07T15:04:12  *** bhasker <bhasker!~bhasker@64.9.243.180> has quit IRC (Ping timeout: 264 seconds)
2010-11-07T15:06:16  <Migi32> I'm bored. I think I'm going to make a "trailer" to advertise PlanetWars on YouTube a bit. Does anyone remember any cool moments (preferably not longer than like 15 seconds) in games that would fit in such a video?
2010-11-07T15:07:03  <Migi32> I know that the average Youtuber probably won't be interested, but meh
2010-11-07T15:09:41  *** Utkarsh <Utkarsh!~admin@117.201.80.110> has quit IRC (Ping timeout: 245 seconds)
2010-11-07T15:10:15  *** Accoun <Accoun!~1@static-213-88-2-78.chebnet.ru> has joined #aichallenge
2010-11-07T15:14:36  *** perror <perror!~fleury@aldebaran.labri.fr> has quit IRC (Quit: Bye all !)
2010-11-07T15:15:43  <contestbot> Top 10 players: bocsimacko(4371), shangas(4238), bix0r4ever(4187), rsergio(4088), asavis(4067), george(4063), dmj111(4058), Xrillo(4034), GreenTea(4032), Accoun(4013)
2010-11-07T15:15:52  *** Utkarsh <Utkarsh!~admin@117.201.80.218> has joined #aichallenge
2010-11-07T15:19:20  *** hellman <hellman!~hellman@nat232-248-205-109.tvoe.tv> has quit IRC (Remote host closed the connection)
2010-11-07T15:21:23  <demerzel_> nice i can handle a 3 ply brute force tree in under a second - run out of stack on 4 haha
2010-11-07T15:25:14  <Error323> goddamnit
2010-11-07T15:25:23  <Error323> shit is failing here >_>
2010-11-07T15:27:39  *** sigh <sigh!~sigh@115.114-50-210.dynamic.dsl.mel.iprimus.net.au> has joined #aichallenge
2010-11-07T15:28:11  *** TravisD <TravisD!~Adium@uwsclient-159-148.uws.ualberta.ca> has left #aichallenge
2010-11-07T15:30:01  *** bhasker_ <bhasker_!~bhasker@69-170-40-2.static-ip.telepacific.net> has quit IRC (Quit: bhasker_)
2010-11-07T15:31:23  <Migi32> demerzel_, based on what do you prune?
2010-11-07T15:31:29  <demerzel_> nothing at the moment haha
2010-11-07T15:31:38  <Migi32> is that even possible?
2010-11-07T15:31:42  <demerzel_> that's why i said brute force
2010-11-07T15:31:57  <javagamer> lol, how well does it play?
2010-11-07T15:32:05  <demerzel_> it doesn't it's all test shit atm
2010-11-07T15:32:10  <javagamer> ah
2010-11-07T15:32:26  <Migi32> I mean, sending either 1 or 0 ships to a planet on the first turn is 2^22 possibilities. There has to be some kind of pruning.
2010-11-07T15:32:28  <demerzel_> a move is not 1-n ships to 0-22 planet tho
2010-11-07T15:32:31  <demerzel_> it's a defined capture/defend
2010-11-07T15:32:52  <Error323> it'll be extremely suboptimal
2010-11-07T15:32:57  <Error323> cuz i've tried that ^_^
2010-11-07T15:33:04  <Migi32> ah yes that's whan I meant with pruning.
2010-11-07T15:33:21  <demerzel_> i end up with ~6k at lvl3
2010-11-07T15:34:09  <demerzel_> it is doing all the turn stepping etc and evaluation position for those 6k tho
2010-11-07T15:34:11  <Error323> demerzel_: but still getting 3 ply is impressive
2010-11-07T15:34:26  <Error323> assuming that's under a second
2010-11-07T15:34:26  *** boegel <boegel!~boegel@91.182.237.170> has joined #aichallenge
2010-11-07T15:35:02  <demerzel_> well this is my first foray into 2+ ply but it's all in C and pretty well optimized with that in mind
2010-11-07T15:35:13  <Migi32> that's a branching factor of about 18. Nice.
2010-11-07T15:35:47  <demerzel_> it's too high because it's counting neutrals i might capture as potential sources for ones that i don't want yet
2010-11-07T15:35:58  <demerzel_> ie i capture 4 and 0 becomes an option
2010-11-07T15:36:06  <demerzel_> it shouldn't be ;)
2010-11-07T15:36:37  <Migi32> wait, 3 plies is 3 turns or 1.5 turns?
2010-11-07T15:37:06  <Error323> Migi32: the branching factor is much higher right
2010-11-07T15:37:08  <demerzel_> 3 plies is n turns
2010-11-07T15:37:21  <demerzel_> because it's not limited to just this turn
2010-11-07T15:37:28  <Error323> 3 plies is n turns??
2010-11-07T15:37:37  <Error323> what is n
2010-11-07T15:37:38  <Migi32> I'm not following
2010-11-07T15:38:00  <demerzel_> i might be able to capture 4 planets this turn - it considers all 4 of those in sequence - i might not be able to capture another one for 8 more turns, it will look from there
2010-11-07T15:38:04  <Migi32> but anyway 18*18*18 = 5832 ~= 6000 so yes, it's a branching factor of 18/ply
2010-11-07T15:38:18  <Error323> ohh
2010-11-07T15:38:20  <Error323> like that
2010-11-07T15:38:27  <demerzel_> a move is not turn-dependent is what i'm saying lol
2010-11-07T15:38:44  <Migi32> ah ok.
2010-11-07T15:38:54  <demerzel_> or i guess it is but that's not a limitation on the branching
2010-11-07T15:39:26  <Migi32> so it means you won't be sending ships to more than 3 planets per turn?
2010-11-07T15:39:59  <demerzel_> potentially but i doubt i'll stick with brute force
2010-11-07T15:40:08  <demerzel_> it's evaluating an entire gamestate but it's pure greed atm
2010-11-07T15:40:36  <demerzel_> although the way it's set up is not a problem for evaluating enemy i just don't know what to do with that yet lol
2010-11-07T15:41:13  <demerzel_> i'm kinda excited, never played with this kind of shit before lol
2010-11-07T15:42:38  *** tapwater_ <tapwater_!~tapwater@cpe-70-112-195-55.austin.res.rr.com> has joined #aichallenge
2010-11-07T15:44:21  *** tapwater <tapwater!~tapwater@cpe-70-112-195-55.austin.res.rr.com> has quit IRC (Ping timeout: 240 seconds)
2010-11-07T15:44:22  *** tapwater_ is now known as tapwater
2010-11-07T15:44:59  <demerzel_> hmm how would i implement a/b on pure greed i guess that would just be alpha pruning lol
2010-11-07T15:54:57  *** sigh <sigh!~sigh@115.114-50-210.dynamic.dsl.mel.iprimus.net.au> has quit IRC (Remote host closed the connection)
2010-11-07T15:58:47  <Migi32> by the way I just calculated it: there are 5 034 380 962 779 878 417 648 226 different first moves. Good luck brute-forcing that :P
2010-11-07T15:59:27  *** jaspervdj <jaspervdj!~jaspervdj@zeus.ugent.be> has quit IRC (Quit: NEVER GONNA GIVE YOU UP NEVER GONNA LET YOU DOWN)
2010-11-07T16:01:04  <demerzel_> made up of 2200 separate moves of which all but probably 5-10 can be remove
2010-11-07T16:02:12  <Migi32> all but 1 can be removed, the trick is knowing which can be removed and which can't
2010-11-07T16:02:27  <javagamer> heh
2010-11-07T16:03:11  <Bobng> you could prune that total down a considerable amount
2010-11-07T16:03:12  <demerzel_> yes but you only need to consider 5-10
2010-11-07T16:04:35  <Migi32> I'm really not so sure of that. Most of us are sending just n+1 ships to the neutrals but when you want to play optimally you also have to know what to do with those 10-ish remaining ships.
2010-11-07T16:04:56  <demerzel_> putting that into the tree is highly inefficient
2010-11-07T16:05:11  <demerzel_> although obv an option
2010-11-07T16:05:33  <boegel> Migi32: you could save them for the next turn, when you might really need them
2010-11-07T16:06:07  <Migi32> boegel, I already do, since it's the only thing you can reasonably do in 1 second.
2010-11-07T16:07:03  <boegel> Migi32: :)
2010-11-07T16:07:14  <boegel> Migi32: you can try and anticipate enemy moves, and act on it
2010-11-07T16:07:15  <demerzel_> there's another solution but i'm talking too much haha
2010-11-07T16:07:33  <boegel> Migi32: e.g. send more than n+1 to a neutral planet, to try and make sure you keep it
2010-11-07T16:08:15  <Migi32> I had one bot who just sent all his unused ships to the closest enemy planet and it did surprisingly well on close-distance maps
2010-11-07T16:15:43  <contestbot> Top 10 players: bocsimacko(4370), shangas(4232), bix0r4ever(4156), rsergio(4078), dmj111(4061), Xrillo(4055), george(4049), asavis(4039), GreenTea(4011), Accoun(4004)
2010-11-07T16:37:21  *** adhoc <adhoc!~adhoc@ppp121-45-60-60.lns20.adl2.internode.on.net> has quit IRC (Ping timeout: 240 seconds)
2010-11-07T16:38:57  *** p4p4 <p4p4!~chatzilla@82.113.121.50> has quit IRC (Remote host closed the connection)
2010-11-07T16:42:44  *** boegel <boegel!~boegel@91.182.237.170> has quit IRC (Quit: Leaving)
2010-11-07T16:50:59  *** RainCT <RainCT!~RainCT@ubuntu/member/pdpc.student.rainct> has quit IRC (Remote host closed the connection)
2010-11-07T16:59:31  *** alch <alch!~alch@188.200.98.21> has quit IRC (Remote host closed the connection)
2010-11-07T17:00:32  *** sigh <sigh!~sigh@lithium.mailguard.com.au> has joined #aichallenge
2010-11-07T17:09:37  <Migi32> is the winner decided by the final ship count on turn 200 after the battle phase or the growth phase?
2010-11-07T17:11:28  *** irchs <irchs!~Jan@2002:5ec2:21a8:0:226:bbff:fe1d:f630> has quit IRC (Quit: irchs)
2010-11-07T17:11:39  <Migi32> hmm, reading the specs, battle phase seems most likely
2010-11-07T17:12:28  *** Bobng <Bobng!~T@78.148.53.228> has quit IRC (Ping timeout: 272 seconds)
2010-11-07T17:13:20  <demerzel_> yes
2010-11-07T17:15:44  <contestbot> Top 10 players: bocsimacko(4372), shangas(4230), bix0r4ever(4156), dmj111(4060), rsergio(4060), george(4055), asavis(4042), Xrillo(4031), Accoun(4020), GreenTea(4008)
2010-11-07T17:46:10  *** JamesGecko <JamesGecko!~James@adsl-145-139-34.asm.bellsouth.net> has left #aichallenge ("Leaving")
2010-11-07T17:57:53  *** irchs <irchs!~Jan@94-194-33-168.zone8.bethere.co.uk> has joined #aichallenge
2010-11-07T18:00:17  <irchs> What version of mono is on the servers?
2010-11-07T18:00:55  *** albert_ <albert_!~albert@CPE0014d1c9243c-CM001bd71cede2.cpe.net.cable.rogers.com> has joined #aichallenge
2010-11-07T18:02:46  <demerzel_> whatever one allows 3.0 now iirc
2010-11-07T18:03:02  <demerzel_> oh wait it's 2.4.4
2010-11-07T18:03:13  <demerzel_> unless janzert took my advice on a parallel install
2010-11-07T18:15:44  <contestbot> Top 10 players: bocsimacko(4338), shangas(4196), bix0r4ever(4143), dmj111(4045), rsergio(4041), george(4034), GreenTea(4020), Xrillo(3998), asavis(3998), Hazard(3987)
2010-11-07T18:18:41  *** irchs <irchs!~Jan@94-194-33-168.zone8.bethere.co.uk> has quit IRC (Quit: irchs)
2010-11-07T18:26:41  *** tapwater_ <tapwater_!~tapwater@cpe-70-112-195-55.austin.res.rr.com> has joined #aichallenge
2010-11-07T18:28:42  *** tapwater <tapwater!~tapwater@cpe-70-112-195-55.austin.res.rr.com> has quit IRC (Ping timeout: 255 seconds)
2010-11-07T18:28:43  *** tapwater_ is now known as tapwater
2010-11-07T18:29:25  *** bhasker <bhasker!~bhasker@adsl-76-254-15-251.dsl.pltn13.sbcglobal.net> has joined #aichallenge
2010-11-07T18:30:34  *** Accoun <Accoun!~1@static-213-88-2-78.chebnet.ru> has quit IRC (Ping timeout: 240 seconds)
2010-11-07T18:34:51  *** nann <nann!~apophis@80-219-211-230.dclient.hispeed.ch> has quit IRC (Ping timeout: 255 seconds)
2010-11-07T18:36:44  *** fawek <fawek!~faw@chello089079023212.chello.pl> has quit IRC (Read error: Connection reset by peer)
2010-11-07T18:37:20  <nullkuhl> my games gets deleted from non official server ?
2010-11-07T18:40:36  *** tapwater_ <tapwater_!~tapwater@cpe-70-112-195-55.austin.res.rr.com> has joined #aichallenge
2010-11-07T18:41:05  *** mceier <mceier!~mceier@chello089077208118.chello.pl> has quit IRC (Quit: leaving)
2010-11-07T18:42:19  <demerzel_> yea - rather rapidly at that
2010-11-07T18:42:34  *** tapwater <tapwater!~tapwater@cpe-70-112-195-55.austin.res.rr.com> has quit IRC (Ping timeout: 245 seconds)
2010-11-07T18:42:34  *** tapwater_ is now known as tapwater
2010-11-07T18:45:51  *** tapwater_ <tapwater_!~tapwater@cpe-70-112-195-55.austin.res.rr.com> has joined #aichallenge
2010-11-07T18:46:06  <Zoost> servers seem to be dying :<
2010-11-07T18:46:41  *** tapwater <tapwater!~tapwater@cpe-70-112-195-55.austin.res.rr.com> has quit IRC (Ping timeout: 240 seconds)
2010-11-07T18:46:41  *** tapwater_ is now known as tapwater
2010-11-07T19:04:29  *** Migi32 <Migi32!~migi@d54C1291D.access.telenet.be> has quit IRC (Quit: http://boinc.berkeley.edu)
2010-11-07T19:14:10  *** sigh_ <sigh_!~sigh@lithium.mailguard.com.au> has joined #aichallenge
2010-11-07T19:14:11  *** sigh <sigh!~sigh@lithium.mailguard.com.au> has quit IRC (Read error: Connection reset by peer)
2010-11-07T19:15:45  <contestbot> Top 10 players: bocsimacko(4339), shangas(4214), bix0r4ever(4134), george(4071), rsergio(4052), dmj111(4038), GreenTea(4030), Xrillo(4013), Hazard(4001), asavis(3999)
2010-11-07T19:36:31  *** wh1teside <wh1teside!~admin@adsl-70-131-50-177.dsl.emhril.sbcglobal.net> has joined #aichallenge
2010-11-07T19:37:28  *** Utkarsh <Utkarsh!~admin@117.201.80.218> has quit IRC (Ping timeout: 260 seconds)
2010-11-07T19:39:03  *** golf <golf!47bebb38@gateway/web/freenode/ip.71.190.187.56> has joined #aichallenge
2010-11-07T19:40:34  *** gofl132 <gofl132!47bebb38@gateway/web/freenode/ip.71.190.187.56> has joined #aichallenge
2010-11-07T19:40:40  <lavalamp_> who does hightime? we just had an interesting match: http://72.44.46.68/canvas?game_id=901606
2010-11-07T19:41:40  *** Utkarsh <Utkarsh!~admin@117.201.82.143> has joined #aichallenge
2010-11-07T19:41:44  <lavalamp_> man I wish my bot would be able to hold the 370 Elo it's got on the tcp server atm
2010-11-07T19:43:01  *** Thyker <Thyker!~Thyker@129.2.129.102> has joined #aichallenge
2010-11-07T19:44:02  *** virdo <virdo!~virdoo@bas9-london14-1279330399.dsl.bell.ca> has joined #aichallenge
2010-11-07T19:44:20  *** flag <flag!~root@69-165-173-172.dsl.teksavvy.com> has quit IRC (Ping timeout: 252 seconds)
2010-11-07T19:57:09  *** flag <flag!~root@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge
2010-11-07T20:02:09  *** Utkarsh <Utkarsh!~admin@117.201.82.143> has quit IRC (Ping timeout: 255 seconds)
2010-11-07T20:06:00  *** Cold-Phoenix <Cold-Phoenix!~Cold-Phoe@87.112.90.54.plusnet.ptn-ag2.dyn.plus.net> has quit IRC ()
2010-11-07T20:07:25  <Prillicy> I'm having trouble debugging mybot because I don't know java so I don't know what part of the bot is sending illegal commands to the server
2010-11-07T20:07:28  *** Utkarsh <Utkarsh!~admin@117.201.83.157> has joined #aichallenge
2010-11-07T20:07:40  <Prillicy> does anyone know how to find out what your bot is doing wrong?
2010-11-07T20:14:09  <demerzel_> i think we all do
2010-11-07T20:14:58  <demerzel_> you don't know java but you are debugging a java bot?
2010-11-07T20:14:59  <demerzel_> why?
2010-11-07T20:15:14  *** Appleman1234 <Appleman1234!~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au> has quit IRC (Ping timeout: 240 seconds)
2010-11-07T20:15:30  *** Utkarsh <Utkarsh!~admin@117.201.83.157> has quit IRC (Ping timeout: 272 seconds)
2010-11-07T20:15:45  <contestbot> Top 10 players: bocsimacko(4305), shangas(4192), bix0r4ever(4128), george(4058), rsergio(4043), dmj111(4013), GreenTea(4005), Xrillo(3995), asavis(3986), Accoun(3979)
2010-11-07T20:16:05  <Prillicy> demerzel_: the visualizer/gamme runner is in java
2010-11-07T20:16:15  <Prillicy> I can't figure out why my bot keeps crashing on the gamer
2010-11-07T20:16:27  <demerzel_> why would you need to debug java to find that out?  isnt your bot C#
2010-11-07T20:16:40  <demerzel_> just have it log to a file or something
2010-11-07T20:16:45  <Prillicy> ?
2010-11-07T20:16:55  <Prillicy> sorry, still new to c#
2010-11-07T20:17:00  <Prillicy> never built a log  file before
2010-11-07T20:17:03  <demerzel_> i have no idea how your brain works you ask really bizarre questions lol
2010-11-07T20:17:20  <demerzel_> log file is just putting a log to a file
2010-11-07T20:17:23  <demerzel_> that's literally it
2010-11-07T20:17:28  <demerzel_> it's not a hard-defined thing
2010-11-07T20:17:45  <lavalamp_> Prillicy, just open a file and write debug info to it.... C# must have some equivilent of fprintf...
2010-11-07T20:18:04  <Prillicy> k, I'll research
2010-11-07T20:18:29  <lavalamp_> ...and don't forget to disable that code before submitting to the server :)
2010-11-07T20:19:00  <demerzel_> it's like 3 lines - 1 if you keep a static object -- http://www.csharp-station.com/HowTo/ReadWriteTextFile.aspx
2010-11-07T20:19:12  <demerzel_> textwrite.open, textwriter.write textwriter.close
2010-11-07T20:19:24  *** Utkarsh <Utkarsh!~admin@117.201.82.251> has joined #aichallenge
2010-11-07T20:20:36  <demerzel_> here
2010-11-07T20:20:37  <demerzel_> http://code.google.com/p/planetwars-bot/source/browse/trunk/Logger.cs
2010-11-07T20:20:45  <demerzel_> that's the one i used in C#
2010-11-07T20:21:39  <demerzel_> to log you can still use the AppendFormat("ships {0}", ships) just Logger.logger.Appendformat
2010-11-07T20:21:51  <demerzel_> call Logger.doTurn() after every turn
2010-11-07T20:22:07  <demerzel_> or prior to that if your'e crashing ;)
2010-11-07T20:23:11  <demerzel_> sooo close to having a real tree trying not to rush it haha
2010-11-07T20:26:19  *** Utkarsh <Utkarsh!~admin@117.201.82.251> has quit IRC (Read error: Connection reset by peer)
2010-11-07T20:33:43  *** Utkarsh <Utkarsh!~admin@117.201.83.245> has joined #aichallenge
2010-11-07T20:34:20  *** Azrathud <Azrathud!~azrathud@c-67-166-43-227.hsd1.co.comcast.net> has quit IRC (Quit: Meh)
2010-11-07T20:34:29  *** hornairs <hornairs!~hornairs@modemcable004.116-201-24.mc.videotron.ca> has joined #aichallenge
2010-11-07T20:34:36  *** Azrathud <Azrathud!~azrathud@c-67-166-43-227.hsd1.co.comcast.net> has joined #aichallenge
2010-11-07T20:49:49  *** Accoun <Accoun!~1@static-213-88-2-78.chebnet.ru> has joined #aichallenge
2010-11-07T20:52:46  *** Utkarsh <Utkarsh!~admin@117.201.83.245> has quit IRC (Read error: Connection reset by peer)
2010-11-07T21:02:35  *** Utkarsh <Utkarsh!~admin@117.201.84.217> has joined #aichallenge
2010-11-07T21:10:57  *** phreeza_ <phreeza_!~phreeza@f048169182.adsl.alicedsl.de> has joined #aichallenge
2010-11-07T21:11:33  *** phreeza <phreeza!~phreeza@f054211196.adsl.alicedsl.de> has quit IRC (Read error: Operation timed out)
2010-11-07T21:11:34  *** Utkarsh <Utkarsh!~admin@117.201.84.217> has quit IRC (Ping timeout: 240 seconds)
2010-11-07T21:12:13  *** dmj111 <dmj111!~user@c-98-235-91-228.hsd1.pa.comcast.net> has joined #aichallenge
2010-11-07T21:13:03  *** phreeza_ <phreeza_!~phreeza@f048169182.adsl.alicedsl.de> has quit IRC (Client Quit)
2010-11-07T21:15:46  <contestbot> Top 10 players: bocsimacko(4289), shangas(4184), bix0r4ever(4132), rsergio(4038), george(4026), dmj111(3993), GreenTea(3993), asavis(3980), Xrillo(3975), Accoun(3974)
2010-11-07T21:17:16  *** hornairs <hornairs!~hornairs@modemcable004.116-201-24.mc.videotron.ca> has left #aichallenge
2010-11-07T21:17:48  *** Utkarsh <Utkarsh!~admin@117.201.85.61> has joined #aichallenge
2010-11-07T21:28:50  *** Utkarsh <Utkarsh!~admin@117.201.85.61> has quit IRC (Ping timeout: 252 seconds)
2010-11-07T21:33:00  <antimatroid> yay, one exam left
2010-11-07T21:33:06  <antimatroid> pitty its gross pdes
2010-11-07T21:34:18  *** Utkarsh <Utkarsh!~admin@117.201.84.10> has joined #aichallenge
2010-11-07T21:34:58  <brx> uhm. bots execute their turns simultaneously and there is no "first" bot, right?
2010-11-07T21:35:16  <antimatroid> correct
2010-11-07T21:35:18  <brx> good
2010-11-07T21:35:29  <brx> the server is giving a 33 elo advantage to the first bot.
2010-11-07T21:35:41  <antimatroid> the order goes like this
2010-11-07T21:35:41  <antimatroid> new move -> add growth -> deal with incoming fleets -> send new fleets -> new move
2010-11-07T21:35:58  <antimatroid> 33 elo advantage what?
2010-11-07T21:36:04  <brx> which it shouldn't be doing in this case. apparently that is the default with the bayeselo binary in use
2010-11-07T21:36:39  <brx> antimatroid: in chess the white player is estimated to have a 33 elo advantage over black by virtue of moving first.
2010-11-07T21:36:45  <antimatroid> bocsimacko just has a bot much better than everyone else
2010-11-07T21:36:55  <antimatroid> who said it's doing that?
2010-11-07T21:38:25  <dmj111> first moving bot, I think.  like white in chess.
2010-11-07T21:38:29  <lavalamp_> brx I will laugh hard if they copied the rating stuff and didn't change that
2010-11-07T21:39:04  <lavalamp_> I imagine that could add a fairly significant amount of noise to the ratings
2010-11-07T21:39:19  <lavalamp_> @repo
2010-11-07T21:39:20  <contestbot> lavalamp_: repo = http://code.google.com/p/ai-contest/
2010-11-07T21:41:15  <brx> antimatroid: http://code.google.com/p/ai-contest/source/browse/trunk/planet_wars/backend/readgames.py
2010-11-07T21:41:25  <brx> (sorry, got disconnected)
2010-11-07T21:45:08  <lavalamp_> huh, line 36 here: http://code.google.com/p/ai-contest/source/browse/trunk/planet_wars/backend/process_ratings.py
2010-11-07T21:45:15  <lavalamp_> should that really be an insert?
2010-11-07T21:45:37  <brx> dhartmei seems to do it correctly in his tcp server, however in his server source release he calls "advantage 0" after "mm", this is too late to have effect on the
2010-11-07T21:45:37  <lavalamp_> the database keeps a rating history?
2010-11-07T21:45:41  <brx> ratings.
2010-11-07T21:46:19  <lavalamp_> brx what line in that file should I be looking at?
2010-11-07T21:46:57  <brx> lavalamp_: the one i told antimatroid
2010-11-07T21:47:05  <brx> lavalamp_: end of file.
2010-11-07T21:47:55  <brx> and also http://code.google.com/p/ai-contest/source/browse/trunk/planet_wars/backend/update_rankings.sh
2010-11-07T21:48:26  <lavalamp_> oh I see, it's building a command for bayeselo?
2010-11-07T21:49:34  <brx> yes
2010-11-07T21:49:49  <lavalamp_> ok this would make more sense if I knew how bayeselo took its commands, haha
2010-11-07T21:50:10  <brx> unless they patched the bayeselo executable to set the default advantage to 0, they are missing an "advantage 0" before "mm" there
2010-11-07T21:50:50  <antimatroid> hmmm okay, well i wont shut up about this now until we get a proper answer
2010-11-07T21:50:53  <antimatroid> brx: thanks :)
2010-11-07T21:51:22  <antimatroid> i don't even know who to ask about these problems anymore
2010-11-07T21:51:35  <antimatroid> amstan j3camero jmcarthur: any of you about?
2010-11-07T21:51:42  <amstan> ?
2010-11-07T21:52:01  <antimatroid> there is worry that the rankings never got rid of an advantage to player 1
2010-11-07T21:52:10  <antimatroid> in that in chess the first player has a first mover advantage
2010-11-07T21:52:33  <antimatroid> that would obviously be a really big problem
2010-11-07T21:53:45  <amstan> advantage to player 1? huh?
2010-11-07T21:54:35  <antimatroid> well in chess, when the moves aren't simultaneous, there is considered an advantage (to i think first player) so there is a sort of imbalence to how elo is calculated depending on which player you are
2010-11-07T21:54:55  <antimatroid> obviously one doesn't want that for this, the question is was it ever taken out of the code grabbed to do the rankings
2010-11-07T21:55:56  *** Frontier <Frontier!~Marco_Lei@dslb-088-070-137-096.pools.arcor-ip.net> has joined #aichallenge
2010-11-07T21:58:19  <brx> amstan, antimatroid: http://remi.coulom.free.fr/Bayesian-Elo/#elostat <- "In chess, playing with the white pieces is an advantage estimated to be worth about 33 Elo points. Bayeselo takes this into consideration."
2010-11-07T22:07:50  *** Appleman1234 <Appleman1234!~Appleman1@CPE-60-226-176-19.qld.bigpond.net.au> has joined #aichallenge
2010-11-07T22:15:46  <contestbot> Top 10 players: bocsimacko(4289), shangas(4178), bix0r4ever(4142), george(4029), rsergio(4022), dmj111(3996), GreenTea(3979), Xrillo(3978), asavis(3974), Accoun(3965)
2010-11-07T22:29:02  <demerzel_> hahaha damn
2010-11-07T22:29:04  <demerzel_> nice catch brx
2010-11-07T22:29:43  *** Hi <Hi!407ebe7b@gateway/web/freenode/ip.64.126.190.123> has joined #aichallenge
2010-11-07T22:29:56  <brx> I'm not entirely sure there is a problem. I haven't looked at the bayeselo source the server is actually using. it may have been patched.
2010-11-07T22:30:04  <brx> I'm just wondering.
2010-11-07T22:30:10  *** Hi is now known as Guest67498
2010-11-07T22:32:00  <demerzel_> i highly doubt it
2010-11-07T22:34:04  *** tapwater_ <tapwater_!~tapwater@cpe-70-112-195-55.austin.res.rr.com> has joined #aichallenge
2010-11-07T22:35:20  *** amstan <amstan!~amstan@69-196-139-185.dsl.teksavvy.com> has quit IRC (Read error: Connection reset by peer)
2010-11-07T22:35:28  *** amstan <amstan!~amstan@69-196-169-123.dsl.teksavvy.com> has joined #aichallenge
2010-11-07T22:35:28  *** ChanServ sets mode: +o amstan
2010-11-07T22:35:58  <demerzel_> hmm i apparently misunderstood a/b pruning
2010-11-07T22:36:04  <demerzel_> i was pruning at each node level
2010-11-07T22:36:12  *** tapwater <tapwater!~tapwater@cpe-70-112-195-55.austin.res.rr.com> has quit IRC (Ping timeout: 255 seconds)
2010-11-07T22:36:13  *** tapwater_ is now known as tapwater
2010-11-07T22:36:18  <demerzel_> i really don't understand how a/b helps not have to evaluate all of them
2010-11-07T22:36:25  <demerzel_> assuming the node isn't terminal
2010-11-07T22:36:43  <Prillicy> I think my logs are wrong
2010-11-07T22:37:33  <Prillicy> I'm showing, as I go through the planets, that almost every planet is followed by another planet with the exact same number of ships
2010-11-07T22:37:57  <demerzel_> check your log variables first
2010-11-07T22:38:06  <demerzel_> probably just put Planet(0) in one of them
2010-11-07T22:38:09  <demerzel_> if not yea that's a big prob
2010-11-07T22:38:39  <Prillicy> ?
2010-11-07T22:38:46  <Prillicy> let me show ya
2010-11-07T22:39:11  <Prillicy> now remember, I have a function to sort them based on their distance from the home lanet
2010-11-07T22:39:16  *** yasith <yasith!~yasith@112.135.73.41> has quit IRC (Ping timeout: 272 seconds)
2010-11-07T22:39:21  <Prillicy> actually, let me check that functioin for buts
2010-11-07T22:39:23  <Prillicy> *bugs
2010-11-07T22:40:07  *** gofl132 <gofl132!47bebb38@gateway/web/freenode/ip.71.190.187.56> has quit IRC (Quit: Page closed)
2010-11-07T22:40:08  *** Guest67498 <Guest67498!407ebe7b@gateway/web/freenode/ip.64.126.190.123> has quit IRC (Quit: Page closed)
2010-11-07T22:41:22  <lavalamp_> demerzel_ one of the wikipedia articles on a/b pruning has a nice animated gif that shows how it works
2010-11-07T22:41:45  <demerzel_> i tried to look at it but it is annoying that it's not framed
2010-11-07T22:42:07  <demerzel_> i get the alpha/beta stuff but in every case you are going to the full depth and evaluating?
2010-11-07T22:42:14  <lavalamp_> basically you know what's reasonable from the first branch you evaluate, so anything you find in the next branch that's too good or too bad you can skip
2010-11-07T22:42:32  <lavalamp_> so it only helps in subsequent branches
2010-11-07T22:43:01  <demerzel_> what do you mean 'skip' because you are still evaluating at the full depth
2010-11-07T22:43:10  <demerzel_> to have a value to compare with that is
2010-11-07T22:43:51  <lavalamp_> uhhh
2010-11-07T22:44:43  <lavalamp_> a/b messes with my head, I'm not sure how to answer
2010-11-07T22:44:57  <demerzel_> here's what i did: i kept a running tally of a/b for each node level and only further evaluated down that node if it passed
2010-11-07T22:45:18  <demerzel_> but if i'm understanding 'true' a/b correctly i am evaluating at full depth for every combination regardless, i don't understand what i'm gaining
2010-11-07T22:45:40  <lavalamp_> a/b is garanteed not to miss good branches
2010-11-07T22:45:49  <demerzel_> well yes, it's evaluating every branch
2010-11-07T22:45:50  <lavalamp_> if your moves are ordered well it saves you
2010-11-07T22:46:01  *** Accoun <Accoun!~1@static-213-88-2-78.chebnet.ru> has quit IRC (Ping timeout: 240 seconds)
2010-11-07T22:46:04  <lavalamp_> you shouldn't have to evaluate every branch or something is wrong
2010-11-07T22:46:20  <demerzel_> well not every branch level but every full depth node
2010-11-07T22:46:40  <demerzel_> excluding terminals
2010-11-07T22:46:56  <demerzel_> that's the only thing i see as truly allowing me to 'skip' anythin
2010-11-07T22:47:02  *** Rubicon-|-Cross <Rubicon-|-Cross!~RubiconCr@adsl-99-30-224-80.dsl.pltn13.sbcglobal.net> has joined #aichallenge
2010-11-07T22:47:15  <lavalamp_> It should be allowing you to skip entire branches
2010-11-07T22:47:35  <demerzel_> it is very likely i am misunderstanding but how would you have a value to compare to alpha/beta without evaluating it?
2010-11-07T22:48:04  <lavalamp_> unfortunately I don't know it well enough to formulate a coherent answer :/
2010-11-07T22:48:32  *** yasith <yasith!~yasith@112.135.78.91> has joined #aichallenge
2010-11-07T22:49:40  <demerzel_> i mean calc'ing a/b on each /level/ lets you skip a ton of branches but that's not what it is according to the stuff i've read
2010-11-07T22:49:45  <Prillicy> alright, well the good news is, I think I've worked out almost all of the bugs with my regular code, but now the problem exists with my sort of the list of planets
2010-11-07T22:50:01  <Prillicy> if anybody wants to help me out, here's the comparison function http://pastebin.com/wrGUHYMq
2010-11-07T22:50:11  <Prillicy> I am probably missing something stupid
2010-11-07T22:51:20  *** Utkarsh <Utkarsh!~admin@117.201.84.10> has quit IRC (Ping timeout: 252 seconds)
2010-11-07T22:51:26  <demerzel_> that looks ok to me probably how you're using it
2010-11-07T22:51:59  <Prillicy> pw.NeutralPlanets().Sort(judgeDistN);
2010-11-07T22:52:08  <demerzel_> btw you can just implement IComparable<Planet> and override CompareTo<Planet,Planet> then you can just use List.Sort
2010-11-07T22:52:21  <demerzel_> oh you're using it like a predicate
2010-11-07T22:52:26  <demerzel_> that works
2010-11-07T22:52:38  <Prillicy> but it doesn't, my logs prove that
2010-11-07T22:53:10  <demerzel_> you could turn that whole thing into one line
2010-11-07T22:53:28  <demerzel_> return pw.Distance(current.PlanetID(), x.PlanetID()) - pw.Distance(current.PlanetID(), y.PlanetID());
2010-11-07T22:53:38  <demerzel_> your current.PlanetID is what's messing it up then if you're using it as a predicate
2010-11-07T22:54:00  <demerzel_> it doesn't care what the actual value is on a predicate, only if it's negative, positive or 0
2010-11-07T22:54:13  <Prillicy> http://pastebin.com/Cmgkq0yS <---the log of looping through the array of Planets after a sort
2010-11-07T22:54:22  <lavalamp_> demerzel_, I thought a/b did change at each level
2010-11-07T22:54:27  <Prillicy> neat
2010-11-07T22:54:35  <demerzel_> show me how you're using it Pricilly
2010-11-07T22:54:47  <Prillicy> sure, I'll put up the file
2010-11-07T22:54:59  <demerzel_> make currentPlanetID a static variable, set it, then call sort
2010-11-07T22:55:02  <demerzel_> it should work
2010-11-07T22:55:26  <Prillicy> http://pastebin.com/LqSW5Q3x
2010-11-07T22:55:34  *** albert_ <albert_!~albert@CPE0014d1c9243c-CM001bd71cede2.cpe.net.cable.rogers.com> has quit IRC (Quit: Leaving)
2010-11-07T22:55:50  <Prillicy> that's all of the code from the starter pack that I've changed so far
2010-11-07T22:55:51  <lavalamp_> I was excited to do a/b in go because I can actually use α/β as variable names... :/
2010-11-07T22:55:58  <Prillicy> have yet to get this working successfully
2010-11-07T22:56:25  *** Utkarsh <Utkarsh!~admin@117.201.82.53> has joined #aichallenge
2010-11-07T22:57:31  <demerzel_> Pricilly: you are sorting a return value and then doing nothing with it
2010-11-07T22:57:43  <demerzel_> pw.NeutralPlanets().Sort(more.judgeDistN);
2010-11-07T22:57:52  <demerzel_> you need to create a local var holding that return value then sort it
2010-11-07T22:57:57  <demerzel_> as it is it will sort it then destroy it
2010-11-07T22:58:45  <demerzel_> your next call to NeutralPlanets will return the exact same value every time
2010-11-07T23:00:02  <lavalamp_> one of these days my framework will be perfect
2010-11-07T23:00:19  <Prillicy> oh....
2010-11-07T23:01:10  <demerzel_> lavalamp: i doubt it's an eval on every level, since in just a 3-ply tree the results varied wildly
2010-11-07T23:01:28  <demerzel_> you can have a suboptimal first move that creates an alpha 2nd moe
2010-11-07T23:02:53  <demerzel_> i am sure smarter people than i have noticed that and doubt they are okay with it lol
2010-11-07T23:03:12  <lavalamp_> no you don't do an eval on every level
2010-11-07T23:03:13  <demerzel_> the only advantage i can see with a/b is that you don't evaluate on every level
2010-11-07T23:03:52  <demerzel_> and obv if you find a terminal node you stop but i think any sane programmer would put that in
2010-11-07T23:04:24  *** Mathnerd314 <Mathnerd314!~mathnerd3@dsl.6.28.53.206.cos.dyn.pcisys.net> has joined #aichallenge
2010-11-07T23:04:42  <lavalamp_> maybe it is: if you find one terminal that's too good or too bad you can skip the rest of the terminals from that node
2010-11-07T23:05:04  <demerzel_> but then you're evaluating at every level like i was
2010-11-07T23:05:18  <demerzel_> since you'd have to have a value to compare with ;)
2010-11-07T23:05:26  <lavalamp_> ah no
2010-11-07T23:05:40  <lavalamp_> because a/b values propagate back from the terminal evals
2010-11-07T23:06:33  <demerzel_> here's what i did - given a 3-ply search, move 1 yields 20, move 2 yields 40, move 3 yields 60 (cumulative scores) - if at next branch move 1 yielded <20, abandon branch
2010-11-07T23:06:52  <demerzel_> what i understand ab to do is go to 3 at every combination
2010-11-07T23:07:00  <demerzel_> eval there
2010-11-07T23:07:39  <lavalamp_> yeah you have to descend all the way before evaluating
2010-11-07T23:08:05  <lavalamp_> you can't compare eval at level 1 with an eval at level 2
2010-11-07T23:08:37  <lavalamp_> btw are your plies a move from each player or just a move from a single player?
2010-11-07T23:08:51  <demerzel_> it's negamax
2010-11-07T23:09:00  <demerzel_> easier on me ;)
2010-11-07T23:09:22  <lavalamp_> sure but that's not why you can't compare
2010-11-07T23:09:36  <demerzel_> oh well it's a move from both :P
2010-11-07T23:09:42  <lavalamp_> oh ok
2010-11-07T23:09:43  <lavalamp_> hm
2010-11-07T23:09:53  *** Accoun <Accoun!~1@static-213-88-2-78.chebnet.ru> has joined #aichallenge
2010-11-07T23:10:11  <lavalamp_> I'm really not sure how to do a/b like that; I switched to recursive payoff tables for my bot
2010-11-07T23:10:33  <demerzel_> well even in a pure greed implementation (as i have up and running now) i'm not sure
2010-11-07T23:11:01  <antimatroid> demerzel_: there's a good animation page, hang on
2010-11-07T23:11:24  <demerzel_> i know that one
2010-11-07T23:11:26  <antimatroid> http://wolfey.110mb.com/GameVisual/launch.php?agent=2
2010-11-07T23:11:30  <demerzel_> i hate it because i can't read it before it changes
2010-11-07T23:11:36  <demerzel_> oh good lol
2010-11-07T23:11:49  <demerzel_> thought you meant the wiki animated gif
2010-11-07T23:11:56  <antimatroid> nope
2010-11-07T23:12:01  <antimatroid> i saw that comment ;)
2010-11-07T23:12:18  <demerzel_> ohhhh i get it
2010-11-07T23:12:21  <antimatroid> i'm not convinced on using ab pruning on this
2010-11-07T23:12:26  <demerzel_> because you're returning the terminal values back down teh branch
2010-11-07T23:12:31  <demerzel_> damn that's all it took ty
2010-11-07T23:12:49  <antimatroid> np
2010-11-07T23:13:44  <antimatroid> well sure for evaluating it's useful, but i think that's pretty much free compared to generating
2010-11-07T23:13:53  <antimatroid> and it's using it for iterative deepening that i''m unsure of
2010-11-07T23:14:04  <antimatroid> i don't trust the payoffs to not switch around
2010-11-07T23:14:06  *** yasith <yasith!~yasith@112.135.78.91> has quit IRC (Ping timeout: 272 seconds)
2010-11-07T23:15:48  <contestbot> Top 10 players: bocsimacko(4287), shangas(4180), bix0r4ever(4131), george(4029), rsergio(4019), GreenTea(3986), dmj111(3982), Accoun(3979), Xrillo(3978), fglider(3975)
2010-11-07T23:16:13  <lavalamp_> I said that :P <lavalamp_> because a/b values propagate back from the terminal evals
2010-11-07T23:18:20  *** demerzel__ <demerzel__!488039c1@gateway/web/freenode/ip.72.128.57.193> has joined #aichallenge
2010-11-07T23:18:29  <demerzel__> lol that page crashed ff
2010-11-07T23:19:21  *** demerzel_ <demerzel_!488039c1@gateway/web/freenode/ip.72.128.57.193> has quit IRC (Ping timeout: 265 seconds)
2010-11-07T23:20:20  <demerzel__> nick _demerzel
2010-11-07T23:20:23  *** demerzel__ is now known as _demerzel
2010-11-07T23:21:09  <Prillicy> _demerzel: http://msdn.microsoft.com/en-us/library/b0zbh7b6.aspx according to that, you don't have to assign the sort, it should sort the original variable
2010-11-07T23:21:25  <Prillicy> regardless, when I do sort, I'm getting an error
2010-11-07T23:21:36  <Prillicy> at least, doing it your way
2010-11-07T23:24:07  <Prillicy> cannot convert void to list
2010-11-07T23:24:20  <Prillicy> should I return a list with my function? hmm
2010-11-07T23:24:35  <_demerzel> assign the sort?
2010-11-07T23:24:42  <_demerzel> it's like this
2010-11-07T23:24:51  <_demerzel> you call the function it returns a value, you are inline calling sort on it
2010-11-07T23:24:55  <_demerzel> then it poof - disappears
2010-11-07T23:25:03  <_demerzel> because you didnt' assign it to anything
2010-11-07T23:25:09  <_demerzel> next time you call it you get the normal function value, as expected
2010-11-07T23:25:31  <Prillicy> I don't understand
2010-11-07T23:25:38  <_demerzel> i don't know how else to explain it
2010-11-07T23:25:41  <Prillicy> no, not that
2010-11-07T23:25:49  <Prillicy> when I made the change you suggested
2010-11-07T23:25:52  <Prillicy> it results in an error
2010-11-07T23:25:58  <_demerzel> which change
2010-11-07T23:26:04  <Prillicy> cannot convert type void to type List<Planet>
2010-11-07T23:26:05  <_demerzel> assigning it locally?
2010-11-07T23:26:09  <Prillicy> yeah
2010-11-07T23:26:24  <_demerzel> type void
2010-11-07T23:26:28  <_demerzel> wtf how is starter pack set up
2010-11-07T23:26:33  <_demerzel> hold on lemme see if i still have it
2010-11-07T23:26:43  *** yasith <yasith!~yasith@112.135.85.187> has joined #aichallenge
2010-11-07T23:26:58  <Prillicy> as far as I can tell, list.sort() is an extension method
2010-11-07T23:27:03  <_demerzel> no it's what i thought -- public List<Planet> NeutralPlanets()
2010-11-07T23:27:29  <_demerzel> so you should be able to just List<Planet> localNeutrals = NeutralPlanets().Sort(predicate)
2010-11-07T23:27:50  <Prillicy> http://msdn.microsoft.com/en-us/library/b0zbh7b6.aspx <---msdn doesn't assign the sort either
2010-11-07T23:28:05  <_demerzel> you are misunderstanding the problem
2010-11-07T23:28:10  <Prillicy> ?
2010-11-07T23:28:11  <_demerzel> the Sort is not the problem
2010-11-07T23:28:17  <_demerzel> it probably works as expected
2010-11-07T23:28:40  <Prillicy> you're misinterpreting me. msdn, does list.sort() then foreach loops through the array again
2010-11-07T23:28:42  <Prillicy> *list
2010-11-07T23:28:48  <_demerzel> the problem is you are sorting a return value without assigning it then expecting yoru next function call to return a sorted list
2010-11-07T23:28:52  <Prillicy> no assignment
2010-11-07T23:28:55  <_demerzel> it doesn't work that way
2010-11-07T23:29:05  <Prillicy> so msdn is wrong?
2010-11-07T23:29:20  <_demerzel> urgh
2010-11-07T23:29:52  <lavalamp_> lol
2010-11-07T23:29:58  <lavalamp_> msdn doesn't apply
2010-11-07T23:30:00  <Prillicy> it makes sense, because when I assign it, I get a compiler error that it returns void
2010-11-07T23:30:16  <lavalamp_> you may need to assign it *before* you sort
2010-11-07T23:30:26  <lavalamp_> I don't know c#
2010-11-07T23:30:38  <_demerzel> it's like this: if i have a function like this "int x() { return 1; }" and then say x().Add(1) and expect my next call to x() to return 2 i am in your shoes
2010-11-07T23:31:01  <Prillicy> ah, I see what you mean
2010-11-07T23:31:02  <Prillicy> thanks
2010-11-07T23:31:07  <Prillicy> because of the ()
2010-11-07T23:31:32  <_demerzel> i need to do " int xAdd = x() + 1" then work with xAdd not x()
2010-11-07T23:31:49  <Prillicy> gotcha
2010-11-07T23:32:18  <_demerzel> lavalamp_: no the sorting works he was expecting the next return value to be sorted
2010-11-07T23:32:41  <lavalamp_> but does Sort() return this at the end?
2010-11-07T23:32:52  <Prillicy> the sort doesn't return
2010-11-07T23:32:57  <lavalamp_> (assuming c# has something like a this pointer)
2010-11-07T23:33:07  <_demerzel> correct it's a member function
2010-11-07T23:33:08  <Prillicy> the problem was I needed to assign the () to a nonstatic variable
2010-11-07T23:33:27  <Prillicy> if I got the terminology right
2010-11-07T23:33:29  <_demerzel> the function returns a list which a member function is called on then nothign because it's not assigned
2010-11-07T23:33:49  <Prillicy> right, I'm trying to act on a function, not on a variable
2010-11-07T23:34:02  <_demerzel> the () is a pretty basic tenet of oop programming or i would say programming in general
2010-11-07T23:34:28  <_demerzel> you cant change what a function returns only properties
2010-11-07T23:34:49  <_demerzel> the whole point of oop programming is that you are only dealing with a local copy of whatever without being able to modify it
2010-11-07T23:34:59  <_demerzel> black box ideology
2010-11-07T23:34:59  <Prillicy> again, new to c#
2010-11-07T23:35:19  <_demerzel> () is pretty universal hehe
2010-11-07T23:35:30  <_demerzel> functions vs properties
2010-11-07T23:35:47  <Prillicy> yeah, but I was very confused
2010-11-07T23:39:09  <_demerzel> it's fine not like irc is blowing up atm
2010-11-07T23:39:14  *** Thyker <Thyker!~Thyker@129.2.129.102> has quit IRC (Quit: Leaving)
2010-11-07T23:40:51  *** amriedle <amriedle!~amriedle@CPE0013f7f2be12-CM0013f7f2be0e.cpe.net.cable.rogers.com> has quit IRC (Quit: Lost terminal)
2010-11-07T23:41:01  *** yasith_ <yasith_!~yasith@112.135.107.224> has joined #aichallenge
2010-11-07T23:42:55  *** yasith <yasith!~yasith@112.135.85.187> has quit IRC (Ping timeout: 252 seconds)
2010-11-07T23:46:31  <_demerzel> hmm should i give enemy moves the same consideration i give mine or just +1 lol
2010-11-07T23:46:51  <_demerzel> bah i guess i should