2011-02-12T00:00:11 I only wonder if it's not too generous, but overall yeah it seems a rather nice solution 2011-02-12T00:00:39 oh wait, i was thinking we can turn into them 2011-02-12T00:00:43 what if they all became food :P 2011-02-12T00:00:56 i'd probably go with removing them and setting their score to 0? 2011-02-12T00:01:16 yeah 2011-02-12T00:01:24 i think it should just be an instant lose 2011-02-12T00:01:45 the worry is how much it ends up disrupting the balance of the game for the rest of the players 2011-02-12T00:02:02 *** Mathnerd314 has quit IRC (Ping timeout: 240 seconds) 2011-02-12T00:02:52 e.g. if 4 players in a grid and 1,2;3,4 and 4 crashes out, 2 and 3 seem inordinately benefited by it. so by just freezing 4's ants instead of killing them off it is less of an advantage to 2 and 3 2011-02-12T00:03:13 but yeah, I wouldn't mind if 4 still didn't receive any score 2011-02-12T00:03:23 would seem better actually 2011-02-12T00:04:21 well, just freezing them makes sense, too 2011-02-12T00:04:47 so how about freezing and setting their score to zero? 2011-02-12T00:04:52 yep 2011-02-12T00:04:55 that would seem ideal to me 2011-02-12T00:05:03 i'll throw that into the spec for the moment 2011-02-12T00:07:22 what's interesting is that you could try to detect when an opponent has died and act differently if so 2011-02-12T00:07:38 which leads to the counter strategy of pretending you are dead... 2011-02-12T00:08:46 *** Prillicy_ has joined #aichallenge 2011-02-12T00:08:49 the latter doesn't stand a good chance of working out very well i guess 2011-02-12T00:08:57 what's mcleopolds current scoring system? 2011-02-12T00:09:29 jmcarthur: that's a good point 2011-02-12T00:09:33 it adds more strategy to it 2011-02-12T00:09:38 *** Prillicy has quit IRC (Ping timeout: 240 seconds) 2011-02-12T00:09:40 *** Prillicy_ is now known as Prillicy 2011-02-12T00:23:00 *** tapwater has quit IRC (Quit: tapwater) 2011-02-12T00:26:37 this is what i was envisioning when i was thinking about a mmorts for aichallenge: http://www.goblincamp.com/2011/02/version-v0-14-release/ 2011-02-12T00:28:32 if a bot crashes, it will stop accumulating points 2011-02-12T00:28:53 but, it could still beat a bot that was eliminated before the crash 2011-02-12T00:30:36 McLeopold: i'd set it's score to 0? 2011-02-12T00:30:39 what about a timeout? 2011-02-12T00:30:43 its* 2011-02-12T00:30:51 a timeout is pretty much the same as a crash 2011-02-12T00:30:58 McLeopold: i'm currently doing up a game spec on the wiki 2011-02-12T00:31:07 that;s my point, shouldn't that be considered a loss 2011-02-12T00:31:33 no 2011-02-12T00:32:09 there's no reason to punish a bot that crashes anymore than it not being in the match 2011-02-12T00:32:29 are you removing all their ants? 2011-02-12T00:32:35 or leaving them as "non-moving" enemies? 2011-02-12T00:32:40 no, that would unbalance the game for the others 2011-02-12T00:32:47 non-moving enemies 2011-02-12T00:32:57 they can still claim new ants as well 2011-02-12T00:32:59 sure, but jmcarthur makes a good point that you then want to predict whether an ant is dead 2011-02-12T00:33:36 i think it's the better solution overall though 2011-02-12T00:34:00 what's the current scoring system? 2011-02-12T00:34:20 kills 2011-02-12T00:34:31 plus food bonus at end 2011-02-12T00:35:04 what is jmcarthur's point? 2011-02-12T00:35:48 you want to see if an enemies ants aren't "moving" 2011-02-12T00:36:00 and you might then want to "play dead" 2011-02-12T00:36:41 a kill == having an ant in range of an ant dying from battle yes? and what's the food bonus? 2011-02-12T00:37:32 the food bonus is extra points to the last bot standing, based on the amount of food that would have been spawned 2011-02-12T00:37:57 do they get points for all the food currently on the map as well? 2011-02-12T00:38:02 it is to take away incentive for dragging out the game, and to ensure the last bot wins the score 2011-02-12T00:38:18 I hadn't thought of that 2011-02-12T00:38:20 probably 2011-02-12T00:38:21 yes 2011-02-12T00:38:35 okay, i think i'm good to do the spec 2011-02-12T00:38:45 so, with the input 2011-02-12T00:38:49 we have 2 options 2011-02-12T00:38:58 I don't think the options will change at this point 2011-02-12T00:39:07 i am putting both battle resolutions in there so there's a place with both clearly defined 2011-02-12T00:39:08 we will just pick the best one 2011-02-12T00:39:16 ok 2011-02-12T00:39:33 oh, did you see my comment about attack and birth radius? 2011-02-12T00:39:46 i think attack = 5 and birth = 1 2011-02-12T00:39:56 yeah, those are still up for debate 2011-02-12T00:40:03 so attack is with knight moves 2011-02-12T00:40:23 that gives a nice circle for attack and ensures food squares never actually die 2011-02-12T00:45:08 I like food squares dying 2011-02-12T00:47:27 i don't think it really matters or changes anything 2011-02-12T00:54:33 :O 2011-02-12T00:54:37 food is alive? 2011-02-12T00:54:45 i don't want my ants to eat any now 2011-02-12T00:54:48 that is mean 2011-02-12T00:55:17 :P 2011-02-12T00:55:46 on the other hand i don't care about eating enemy ants nom nom 2011-02-12T00:57:14 McLeopold: what if battle resolution also treated a dying ant like food :\ 2011-02-12T00:57:22 then we really could call it zombie ants 2011-02-12T00:57:54 that would be worse than option 2 2011-02-12T00:58:11 :P 2011-02-12T00:58:26 i don't really want to change anything at this point 2011-02-12T00:58:28 it was more an insight 2011-02-12T00:58:48 other than choosing things, i think the game is pretty well defined 2011-02-12T00:59:05 is it fun and balanced and fair? 2011-02-12T00:59:26 yes, depends on battle option if you ask me :P, yes 2011-02-12T00:59:32 :D 2011-02-12T00:59:40 battle option? 2011-02-12T00:59:45 don't ask :P 2011-02-12T00:59:54 i choose howitzers 2011-02-12T00:59:58 or h bombs :-) 2011-02-12T01:01:14 battle option 1: an ant with more enemy ants in it's fight radius than any of those ants enemies in range dies 2011-02-12T01:01:14 battle option 2: iteratively remove closest enemy ants in fighting range 2011-02-12T01:01:26 *** delt0r_ has joined #aichallenge 2011-02-12T01:02:27 *** delt0r___ has quit IRC (Ping timeout: 240 seconds) 2011-02-12T01:03:29 what about more enemies than friendlys in range. then you can send in the cavalry to support places that are going to be lost :D 2011-02-12T01:03:54 i'm pretty sure that has issues 2011-02-12T01:04:02 i can't remember what they are off the top of my head 2011-02-12T01:04:14 we've pretty much ruled out everything but the above two options 2011-02-12T01:05:45 didn't that one lead to blocking? 2011-02-12T01:07:40 *** Chris_0076 has quit IRC (Ping timeout: 250 seconds) 2011-02-12T01:08:04 hmm i can see how that would happen 2011-02-12T01:09:00 i don't think your battle option 2 makes any sense antimatroid :D 2011-02-12T01:09:22 very random rule :D 2011-02-12T01:09:30 McLeopold: are you taking the last issued command for each ant? or first? or.. 2011-02-12T01:09:48 first 2011-02-12T01:09:54 pyro-: neither do i :P 2011-02-12T01:10:04 McLeopold: i think he's talking about dimkadimons option 2011-02-12T01:10:12 pyro-: I think about the only person that does like option 2 is Jeff 2011-02-12T01:10:14 which led to blocking of arbitrary widths 2011-02-12T01:10:19 if i were making such a game i would have health :-) 2011-02-12T01:10:44 pyro-: i did that, people didn't like it 2011-02-12T01:10:57 health would be nice 2011-02-12T01:11:00 it had home bases too, which you had to try and invade, and take food back there to spawn more ants 2011-02-12T01:11:02 but may be too complicated 2011-02-12T01:11:12 collision = attack. remove one with smaller health. subtract smaller health from bigger health 2011-02-12T01:11:16 you could see it in action, but frontiers laptop is off 2011-02-12T01:11:20 collison with food = more health 2011-02-12T01:11:38 or smt ;) 2011-02-12T01:11:42 every ant can deal 1 damage per turn 2011-02-12T01:12:03 so, it will spread it's damage to all enemies in range 2011-02-12T01:12:12 pyro-: play around with resolution method 1, it's so awesome 2011-02-12T01:12:18 oh that sounds fun too :D 2011-02-12T01:12:25 but your not having health? 2011-02-12T01:12:27 :( 2011-02-12T01:12:29 so, if a single ant comes in, then several ants can take the hit, and save each other 2011-02-12T01:12:50 attack diffusion :D 2011-02-12T01:12:54 it doesn't need it 2011-02-12T01:13:01 but, you'd have to track all your ant's current health 2011-02-12T01:13:24 also you could do spawn new ant = divide up ants health :D 2011-02-12T01:13:42 pyro-: the game is good already 2011-02-12T01:13:46 then you get a choice between mighty single ant or spread out weak ants :D 2011-02-12T01:13:51 people are generally happy with it, and it's well defined 2011-02-12T01:13:55 :D 2011-02-12T01:14:13 if only one person does like your option 2 why not you using option 1? :D 2011-02-12T01:14:22 they're the person with the decision 2011-02-12T01:14:36 and they made the decision already? 2011-02-12T01:14:40 no 2011-02-12T01:14:45 they aren't unreasonable 2011-02-12T01:14:50 he* 2011-02-12T01:15:15 i will start being unreasonable 2011-02-12T01:15:29 that will be fun :D 2011-02-12T01:16:05 antimatroid you said its well defined. may i read about it somehwere? 2011-02-12T01:17:50 i'm writing a spec atm 2011-02-12T01:17:58 oh :D 2011-02-12T01:18:09 then i can already nearly make a starter pack :D 2011-02-12T01:18:29 you will probably want to wait for more decisions to be made 2011-02-12T01:18:33 particularly about input 2011-02-12T01:18:56 input is pretty well defined 2011-02-12T01:19:12 depending on the option? 2011-02-12T01:19:24 with the current option, is that moving to outputting info for all currently seeable squares? 2011-02-12T01:21:14 *** Chris_0076 has joined #aichallenge 2011-02-12T01:21:34 output all food and ant 2011-02-12T01:21:40 output land and wall only once 2011-02-12T01:21:47 output all dead ants each turn 2011-02-12T01:22:01 so, you must clear your list of food 2011-02-12T01:22:34 if you want to remember, it is up to your bot to calc the view distance 2011-02-12T01:23:29 i'm fairly strongly opposed to that :P 2011-02-12T01:23:38 I know 2011-02-12T01:23:59 the specs will have that at least for now 2011-02-12T01:24:12 * janzert doesn't understand why not just outputting one line per map row with non-visible squares just say X'ed out 2011-02-12T01:24:31 janzert: that's the other option 2011-02-12T01:24:32 especially if the bots are already being given the map dimensions 2011-02-12T01:24:59 it seems a lot easier to debug and look at 2011-02-12T01:25:24 and probably easier for the bot to use too 2011-02-12T01:25:25 it would be, but the starter pack would have the ability to turn the input into ascii art if needed 2011-02-12T01:26:17 being able to look at the IO in a text editor and have it make intuitive sense is quite nice 2011-02-12T01:27:20 okay, let's take a vote 2011-02-12T01:27:33 option 1, changes only, option 2, map in ascii 2011-02-12T01:28:46 for a 128x128 map, that's 16,512 bytes per turn 2011-02-12T01:28:52 clearly I'd go for two 2011-02-12T01:29:42 i'm with 2 2011-02-12T01:30:04 even though I think raw data sent is a low priority concern, how many ants/food/etc. is the crossover for option 1 2011-02-12T01:30:34 you can give players the char map of the smallest rectangle containing all visible squares 2011-02-12T01:30:44 and pass the top left and bottom right coords before sending it 2011-02-12T01:31:13 if you're giving them the full dimensions anyway, it seems like extra work that isn't worth it 2011-02-12T01:31:14 2064 changes in state 2011-02-12T01:31:48 janzert: assuming you treat the whole map as unseen at the start of each turn, the difference is reading values to set iterators to and filling in based on that 2011-02-12T01:31:53 antimatroid: if they have ants all over the map, not connected, that is harder than it sounds 2011-02-12T01:32:12 McLeopole: store max/min row/col 2011-02-12T01:32:20 antimatroid: I meant more from the engine perspective 2011-02-12T01:32:26 ah yeah, sure 2011-02-12T01:32:32 but yes, it's minimal extra work either way 2011-02-12T01:32:38 yeah, but vision may wrap 2011-02-12T01:33:00 oh, you're right 2011-02-12T01:33:03 * pyro- just saw the phrase "visible squares" 2011-02-12T01:33:07 there ist fog? :D 2011-02-12T01:33:13 it'd still work, just show too much sometimes 2011-02-12T01:33:18 yes, I posted a movie in the forum 2011-02-12T01:33:30 pyro-: yeah :P 2011-02-12T01:33:46 and it's n-player not 2 player 2011-02-12T01:35:53 i almost think the game could be more interesting without the fog of war with the n-players 2011-02-12T01:36:30 how far can you see? 2011-02-12T01:37:03 * pyro- thinks every game in the competition to date has had perfect information 2011-02-12T01:37:30 i would be very surprised if it changes from sqrt(96) 2011-02-12T01:37:58 that's using the euclidean metric too 2011-02-12T01:38:32 it might 2011-02-12T01:38:39 if we go with a smaller attack radius 2011-02-12T01:39:03 maybe, i still think it's a good view radius 2011-02-12T01:39:11 fog does make it much harder :-) 2011-02-12T01:39:36 i think fog discourages the whole see what everyone else is doing and coopeate etc. 2011-02-12T01:39:41 but that's alright, it's still awesome 2011-02-12T01:40:18 *** bhasker has joined #aichallenge 2011-02-12T01:40:52 im assming that bots will clump there ants in an area that is there controlled territory 2011-02-12T01:41:05 its a bit useless to only know where your own ants are :D 2011-02-12T01:41:12 with option 2, you want to fight single ant battles 2011-02-12T01:41:17 and kamakazee them into the enemy 2011-02-12T01:41:29 which also means there is no balance with also wanting to spread out for finding food 2011-02-12T01:41:43 option 1 gives incentive to clump up, which gives balance to wanting to spread for food 2011-02-12T01:43:13 how will you detirmine whether you want to defend your area or crush a player if you can't see how many ants has that player? 2011-02-12T01:44:31 yes i am asking you to design my bot for me :D 2011-02-12T01:44:46 that's part of the appeal fog does offer 2011-02-12T01:44:54 there is balance from that 2011-02-12T01:44:59 apeal? no it make me sad 2011-02-12T01:45:02 start to cry 2011-02-12T01:45:24 i can only luanch test attacks and see if they die or not. i cant actually make any stratagy :'( 2011-02-12T01:46:46 the rules are set that i think you want to at least fight what you can see 2011-02-12T01:46:52 you wont know how much back up to send 2011-02-12T01:47:05 part of why i like the larger vision radius is because you still get a decent amount of visibility 2011-02-12T01:47:10 particularly once you get more ants 2011-02-12T01:47:49 yes. i can charge at 1 ant becuase i have near him 3 and then find out he has 20 ants just behind one ant and i loose 2011-02-12T01:48:04 then all my ants cry very much 2011-02-12T01:50:03 *** bhasker has quit IRC (Ping timeout: 240 seconds) 2011-02-12T01:50:23 *** bhasker has joined #aichallenge 2011-02-12T01:50:28 then you run backwards 2011-02-12T01:50:34 until your backup arrives 2011-02-12T01:50:46 and hope they can't get you dodging through walls 2011-02-12T01:51:27 the good thing about a smaller attack radius is you don't need as many ants about to still get at least a kill in 2011-02-12T01:51:39 so then stratagy is don't send out single ant ? 2011-02-12T01:51:47 everyone grow out slowly like a mushroom ? 2011-02-12T01:52:17 and you dont know how strong is your neibhour so you must simply keep trying to mushroom out even when you hit him 2011-02-12T01:52:31 becuase if you divert too many forces you may find yourself dead 2011-02-12T01:52:48 if he is stronger 2011-02-12T01:52:54 but you want to spread out so you can see more and get food as it replenishes 2011-02-12T01:52:56 *** bhasker has quit IRC (Read error: Connection reset by peer) 2011-02-12T01:53:05 (oh yeah, there's replenishing food popping up) 2011-02-12T01:53:12 *** bhasker has joined #aichallenge 2011-02-12T01:53:15 yes 2011-02-12T01:53:40 but i have no way to estimate if i get enough food to compensate for attacking him 2011-02-12T01:54:20 i know not what area i can control once he is dead 2011-02-12T01:54:31 nor how much he is capable of defending 2011-02-12T01:54:50 i dont like your fog at all :D 2011-02-12T01:58:20 that is part of the point of fog of war 2011-02-12T01:58:24 you don't have perfect information 2011-02-12T02:00:40 yes 2011-02-12T02:00:48 guessing at things is no fun :'( 2011-02-12T02:04:48 *** Frontier has joined #aichallenge 2011-02-12T02:05:48 pyro-: http://marco-leise.homedns.org/antgui/v3/ 2011-02-12T02:05:56 that was the ant game i made with hp 2011-02-12T02:06:33 Frontier ^^ there did the visualiser 2011-02-12T02:07:29 :) 2011-02-12T02:07:32 no idea what is going on 2011-02-12T02:07:44 hard to follow. but i think it look more fun than the game you have here currently 2011-02-12T02:08:09 if you have food deposited on your home base at the start of a turn, an ant spawns, then you have to go out and harvest, attack and invade their base 2011-02-12T02:08:22 if you invade their base, you win, and there is fog of war 2011-02-12T02:08:24 ok the numbers above the ants are "hp | harvest" 2011-02-12T02:08:30 i don't know, i think the current game will be quite cool 2011-02-12T02:08:55 pff fog is lame ;) 2011-02-12T02:09:05 it means you don't know where their base is 2011-02-12T02:09:09 you have to go and find it 2011-02-12T02:09:32 oh 2011-02-12T02:09:36 the maps are symmertic, but the start base could be anywhere 2011-02-12T02:09:38 at least you have an excuse for it 2011-02-12T02:09:41 :p 2011-02-12T02:09:55 the ants can pass on their harvest to other ants closer to the ant hill 2011-02-12T02:10:03 i dont know what fog supposed to be adding to the game here 2011-02-12T02:10:07 and colliding ants fight to the death 2011-02-12T02:10:14 friend ants colliding fuse and hp just sums 2011-02-12T02:10:21 to your game at least it add "searching" component :D 2011-02-12T02:10:27 pyro-: you don't get told the number of players 2011-02-12T02:10:44 and you have to search out to see what the map looks like 2011-02-12T02:10:45 so its even worse ;) 2011-02-12T02:10:53 that's not worse 2011-02-12T02:10:58 imagine real life colony wars 2011-02-12T02:11:04 they don't know how many other colonies there are 2011-02-12T02:11:06 you have no information to make a global stratagy 2011-02-12T02:11:24 you have a global strategy to maximise vision to have a global strategy 2011-02-12T02:11:26 yo dog 2011-02-12T02:12:27 and the only way to acomplish that stratgy apart froming owning the map by which time you've won anyway, is to send out scouts and weaken your force and die 2011-02-12T02:12:36 so you must be content with looking only at your own ants ;) 2011-02-12T02:13:39 https://github.com/aichallenge/aichallenge/wiki/Ants-Game-Specification 2011-02-12T02:13:45 that's still very much a work in progress 2011-02-12T02:14:24 the bottom is wrong too, i haven't changed it fully from planet wars rules 2011-02-12T02:15:04 the objective is wrong too 2011-02-12T02:15:41 yes 2011-02-12T02:15:53 you talk of killing enemies and then later talk of scoring 2011-02-12T02:16:05 McLeopold: The objective is to have the highest score at the end of the game, which consists of killing the most ants and trying to be the last remaining colony. 2011-02-12T02:16:14 is that sufficient for the very introductory explanation of the game? 2011-02-12T02:16:15 yup 2011-02-12T02:16:24 yup 2011-02-12T02:16:27 "and" 2011-02-12T02:16:34 you must favor fighting 2011-02-12T02:16:41 and being the last bot left 2011-02-12T02:16:47 what happen if a kill the most ants but b is the last one standing? 2011-02-12T02:16:47 being the last remaining colony doesn't ensure you win 2011-02-12T02:17:00 depends on how much the difference is and the food bonus 2011-02-12T02:17:01 surviving longer than someone else by hiding will not get you a good score 2011-02-12T02:17:03 read the scoring part 2011-02-12T02:17:24 pyro-: hopefully you got second place 2011-02-12T02:17:27 McLeopold: it may cause the end to be a cat and mouse game 2011-02-12T02:17:41 or, if he doesn't take over until the very end, you may still win 2011-02-12T02:17:55 antimatroid: how so? 2011-02-12T02:18:10 you want to run around and minimise the food bonus they'll get 2011-02-12T02:18:23 i guess you want to do that for survival anyway 2011-02-12T02:18:25 nevermind 2011-02-12T02:18:34 also. ant from a and b in range of ant from c. c die and both a and b gain one point? 2011-02-12T02:18:52 if a and b are in range of c, they all die 2011-02-12T02:18:58 and 3 points are handed out 2011-02-12T02:20:24 "A bot gets a point for every enemy ant they kill." i cant work out how its decided who was the murderer 2011-02-12T02:20:39 we have a formula 2011-02-12T02:21:21 every ant that dies spreads it's 1 point out to all enemies in range 2011-02-12T02:21:45 normally there will only be 2 ants that fight, so it will be 1 point each 2011-02-12T02:21:54 *** Prillicy_ has joined #aichallenge 2011-02-12T02:21:55 *** Prillicy has quit IRC (Ping timeout: 240 seconds) 2011-02-12T02:22:01 *** Prillicy_ is now known as Prillicy 2011-02-12T02:22:08 if you get lucky and kill 2 ants with 1, you get 2 points, they get 1 2011-02-12T02:22:20 then in my situation above 2011-02-12T02:22:29 1 point each 2011-02-12T02:22:40 each ant give 1/2 point to the other 2 2011-02-12T02:22:44 but i killed 2 ants! 2011-02-12T02:22:53 no, 3 ants died 2011-02-12T02:22:57 when one died i was in rnage. when other died i was in range 2011-02-12T02:23:03 you killed half of each ant 2011-02-12T02:23:49 XD 2011-02-12T02:24:03 the half of ant i didnt kill can still run around :D 2011-02-12T02:24:04 McLeopold: i would have expected 2 points for that 2011-02-12T02:24:20 3 different ants all in range? 2011-02-12T02:24:25 yeah 2011-02-12T02:24:29 each player gets 2 points 2011-02-12T02:24:31 no, only 3 point 2011-02-12T02:24:41 it wouldn't be fair to bots not in the battle 2011-02-12T02:24:41 thats different to your own rule 2011-02-12T02:24:50 so if 2 people kill 1 player, each gets half a point? 2011-02-12T02:24:59 each bot gets 2*1/2 i think 2011-02-12T02:25:03 it say if an ant die all enemies in range get a point 2011-02-12T02:25:09 but thats not what you describe :D 2011-02-12T02:25:16 pyro-: i wrote what that says 2011-02-12T02:25:20 if 2 people kill 1 ant, then there are 3 ants all dying together 2011-02-12T02:25:25 but i only wrote that trying to interpret mcleopolds rule 2011-02-12T02:25:27 atm he is right 2011-02-12T02:25:37 not necessarily 2011-02-12T02:25:42 a..b..c 2011-02-12T02:25:54 oh, you're still with option 2 2011-02-12T02:25:58 yeah 2011-02-12T02:25:58 i'm thinking option 1 always :P 2011-02-12T02:26:01 ok 2011-02-12T02:26:10 such a complecated game 2011-02-12T02:26:20 it's cool complicated 2011-02-12T02:26:31 i do not agree :D 2011-02-12T02:26:39 with that, we can still determine which ants did the kill, and the 1 point will be spread between them 2011-02-12T02:26:43 but i can still try and win it ;D 2011-02-12T02:27:14 McLeopold, if 2 a ants and 1 b ant kill a c ant 2011-02-12T02:27:18 does a get 2/3 points? 2011-02-12T02:27:24 yes 2011-02-12T02:27:34 the 1 point is spread per ant, not per player 2011-02-12T02:27:37 floating point scores :| 2011-02-12T02:28:14 or exact fractions used? 2011-02-12T02:28:15 depending on the attack range, we can just multiply the per ant score to avoid floating point 2011-02-12T02:28:21 hmmm, i prefer my one again :P 2011-02-12T02:29:06 i'll put yet another option split in the spec :P 2011-02-12T02:30:29 i know you wont consider it since your gamnes already made. but its why i suggested health earlier. its more easy to see who attack who else and who is just standing nearby :D 2011-02-12T02:30:30 each ant is worth 840 point 2011-02-12T02:31:10 that is not divisible by 7 is it ^^ 2011-02-12T02:31:14 it is 2011-02-12T02:31:21 wow 2011-02-12T02:31:48 but its not divisible by 9 :D 2011-02-12T02:32:02 or 865 :D 2011-02-12T02:32:04 make that number as large as the max amount of ants in attack range 2011-02-12T02:32:11 no 2011-02-12T02:32:31 yes, that would be for option 1 2011-02-12T02:32:32 you can't get that many ants in the attack range i don't tihnk 2011-02-12T02:32:48 just the perimeter 2011-02-12T02:33:10 not just the preimeter 2011-02-12T02:33:15 the 2 edge perimeter? 2011-02-12T02:33:15 that's what I meant anyway 2011-02-12T02:33:42 :P 2011-02-12T02:33:45 so it was 2011-02-12T02:36:04 *** bhasker has quit IRC (Quit: bhasker) 2011-02-12T02:36:07 scoring would be less intuitive if it's going up 840 at a time 2011-02-12T02:37:03 the visualizer can divide it back out 2011-02-12T02:38:17 and food bonus is *840? 2011-02-12T02:38:25 per food 2011-02-12T02:38:30 McLeopold: About that scoring. Should I do my own calculations or will the replay string include scores for each player? 2011-02-12T02:38:47 Frontier: the go message is followed by scores 2011-02-12T02:38:48 you'll get and end score 2011-02-12T02:38:50 same as end 2011-02-12T02:39:02 no scores during the game 2011-02-12T02:39:17 yes? no? 2011-02-12T02:39:29 McLeopold; he;s talking about the visualiser 2011-02-12T02:39:42 oh 2011-02-12T02:40:01 yeah, whatever you need for the visualizer, we'll give you 2011-02-12T02:40:51 Ok, I guess with a scoring like that it should be in the replay string. Planet Wars was more straight forward there 2011-02-12T02:42:13 since the replay string can see everything, you'll be given the wall info once, then nothing but ant, food and dead ant locations 2011-02-12T02:42:16 along with score 2011-02-12T02:42:17 McLeopold: i'm adding a visualiser input parameter for that 840 thing 2011-02-12T02:42:28 840 is not set in stone 2011-02-12T02:42:43 exactly 2011-02-12T02:42:43 therefor it is a parameter :) 2011-02-12T02:42:50 it's just the lcm of 1-8 2011-02-12T02:42:56 yeah 2011-02-12T02:43:08 that's why i'm making it a parameter :P 2011-02-12T02:43:13 ok 2011-02-12T02:43:47 javascript doesn't have integers anyway... 2011-02-12T02:44:03 well, I thought it has 2011-02-12T02:44:14 only floating point 2011-02-12T02:47:30 McLeopold: does the engine send the bots "go" almost at the same time, or at the end of sending each bot input? and does the timer start after sending go or when sending input starts? 2011-02-12T02:48:00 at the end of each input, but I like your idea 2011-02-12T02:48:17 I'm writing the game class, that should plug into an engine 2011-02-12T02:48:23 so my engine may not be used 2011-02-12T02:48:32 yeah that's fine 2011-02-12T02:48:46 hopefully otherwise the engine will at least be made to the specs 2011-02-12T02:49:32 McLeopold: I found a website mentioning an internal ToInt32 function. And I use to write "number | 0" to trigger that and cast a number to an int 2011-02-12T02:51:10 do you have a link? 2011-02-12T02:51:36 in javascript, number | 0 would just be a way to default to 0 if number was falsy 2011-02-12T02:51:48 is the non existance of a legal move for either player possible? 2011-02-12T02:51:57 http://www.jibbering.com/faq/notes/type-conversion/ 2011-02-12T02:52:09 not really 2011-02-12T02:52:42 you'd have to have all ants in a 1x1 space surrounded by wall 2011-02-12T02:53:04 "number | 0" works very well to trunc() floating point numbers in my visualizer 2011-02-12T02:53:12 what happens if you move an ant into an ant? 2011-02-12T02:53:17 they die 2011-02-12T02:54:26 enemy ants can't land in the same square if the attack radius is >=2 2011-02-12T02:55:06 ok, for a game of 200 turns, sending full map is 485,627 bytes, changes only is 42,295 2011-02-12T02:55:44 so.. char maps? 2011-02-12T02:55:52 yeah 2011-02-12T02:56:11 so, do we want to store 1 meg per game for the replays? 2011-02-12T02:56:34 do we want to have that much bandwidth taken on the web server? 2011-02-12T02:56:40 don't you only need the visualiser stuff? 2011-02-12T02:56:40 janzert: ^^ 2011-02-12T02:56:41 I only need the bare bone version 2011-02-12T02:56:45 why would you not delta encode? 2011-02-12T02:57:12 so they're within about 10x of each other, sounds fine 2011-02-12T02:57:14 some think it is easier to debug and code for if you send the whole map 2011-02-12T02:57:19 encode it whatever way you see fit, but i thought the current format was ok 2011-02-12T02:57:34 janzert: you know the web server better than me 2011-02-12T02:57:37 Frontier: it's the bot input i want to be a char map 2011-02-12T02:57:38 are you sure about that? 2011-02-12T02:57:51 McLeopold: I don't want to change the input for the visualiser 2011-02-12T02:58:02 that is good, and will decrease the size of game strings 2011-02-12T02:58:26 for the replay data, I don't care as much 2011-02-12T02:59:05 in either case it really needs to not be stored in the database like the last contest did 2011-02-12T02:59:05 McLeopold: give me a minute to update that specs page, then see if we all agree 2011-02-12T03:00:03 and it would be best if it can be stored in such a way that it can be served as a static file directly by the webserver 2011-02-12T03:00:04 janzert: So replays would be stored as files? 2011-02-12T03:00:45 janzert: the full map compresses much better, only 2x the size 2011-02-12T03:00:46 Frontier: yeah, I think so although someone had mentioned just storing them in amazon storage 2011-02-12T03:00:52 I currently use a http request in the visualizer to download the replays. the web server is 100% static 2011-02-12T03:01:11 if most everyone uses a browser that requests gzip, we are only going to double bandwidth 2011-02-12T03:01:46 but replays can be cached when there is no parameters involved in the requests 2011-02-12T03:01:47 McLeopold: I rather thought that might indeed be the case, and I do expect most browsers will be accepting gz compressed responses 2011-02-12T03:02:31 Frontier: yep, although the current webserver setup is messed up in that it sends no cache headers for all requests 2011-02-12T03:02:42 okay, I'll look into a modified input scheme where we attempt to send only the visible part to bots 2011-02-12T03:02:43 that should really be changed too :) 2011-02-12T03:03:23 lots of stuff that should get changed :/ 2011-02-12T03:03:29 Frontier: try and get the visualizer to use localStorage 2011-02-12T03:03:42 why? 2011-02-12T03:03:46 or rather what for? 2011-02-12T03:03:52 have a link with information, im not really a web developer? 2011-02-12T03:03:58 so a person can store replays locally if they want 2011-02-12T03:04:24 well, check out dive into html5 2011-02-12T03:04:37 hmm, seems like pretty limited use. Just get the cache headers set right 2011-02-12T03:04:57 McLeopold: what would the use case be? 2011-02-12T03:05:45 if there were different options to set for the visualizer then it seems like localstorage might be handy for that 2011-02-12T03:06:13 but really I think the browser compatibility is still too thin to rely on it 2011-02-12T03:06:20 hmm, yes we talked about zoom levels for that matter 2011-02-12T03:06:41 *** j3camero has quit IRC (Ping timeout: 255 seconds) 2011-02-12T03:10:09 janzert: I don't know if JavaScript is allowed to download from different servers now that I think about it. 2011-02-12T03:10:11 hmm, only about 5% of visitors to the site are using IE, with only about 1% using older than version 8, so maybe you could get away with it 2011-02-12T03:10:41 Frontier: not in general no 2011-02-12T03:11:33 So the replays must be served by the main web server ? 2011-02-12T03:12:49 pretty much, possibly a subdomain works too I'd have to look 2011-02-12T03:12:59 *** amstan has quit IRC (Ping timeout: 276 seconds) 2011-02-12T03:13:21 but I would expect just to have it come from the main server 2011-02-12T03:14:02 alright 2011-02-12T03:16:00 i need to add something to input 2011-02-12T03:16:17 atm it's hard to tell whether you're inputing a char map or the end line 2011-02-12T03:16:43 if we had 14 players, yes it would be 2011-02-12T03:17:28 don't all lines start with a type identifier? 2011-02-12T03:17:49 not for input 2011-02-12T03:18:26 we did have 'M' 2011-02-12T03:18:28 ouch 2011-02-12T03:18:43 i don't like M at the start of every char line 2011-02-12T03:19:04 could I add a T/E line at hte start of each turn 2011-02-12T03:19:40 I think it's really worth it to make each line unambiguously identified by its first word 2011-02-12T03:19:56 okay, i'll do that 2011-02-12T03:20:44 lets people do really easy parsing 2011-02-12T03:21:57 plus, if every line starts 1 char and 1 space, you can make assumptions 2011-02-12T03:22:06 just chop the first 2 chars off each line 2011-02-12T03:22:15 I'd try not to make it 1 char and a space 2011-02-12T03:22:21 we might run out of chars :) 2011-02-12T03:22:37 >:( 2011-02-12T03:22:45 but always followed by a space is certainly good 2011-02-12T03:22:52 antimatroid: how about for bot output 'O' for order 2011-02-12T03:23:37 is the game length given in the pregame setup phase? 2011-02-12T03:23:51 P turns 1000 2011-02-12T03:23:59 yes 2011-02-12T03:24:55 McLeopold: done 2011-02-12T03:25:08 actually, looks a lot like 0 2011-02-12T03:25:14 ok, although I'm not sure why to use instead of just 2011-02-12T03:25:20 W? 2011-02-12T03:25:21 walk 2011-02-12T03:25:31 w = wall 2011-02-12T03:25:36 hmmm 2011-02-12T03:25:37 let's keep em unique 2011-02-12T03:25:41 yeah 2011-02-12T03:25:56 well that's a single char in the spec, so work one out, i've almost finished updating 2011-02-12T03:26:31 that what I'm pushing for, dropping the single char types :P 2011-02-12T03:27:00 just go for the full monte and only use the single char types for things that are used a lot and a natural fit 2011-02-12T03:27:22 like M for map line 2011-02-12T03:27:38 too late, the bike shed is built 2011-02-12T03:27:52 you're using go and ready anyway right? 2011-02-12T03:27:59 dammit 2011-02-12T03:28:26 so not all types are a single char already 2011-02-12T03:28:48 json? 2011-02-12T03:28:57 yaml? 2011-02-12T03:29:04 xml? 2011-02-12T03:29:04 no, just plain :P 2011-02-12T03:29:22 ergh, i need to do formatting hold up 2011-02-12T03:29:56 we could make them strings 2011-02-12T03:30:04 what about bot output, should we bother with the 'O', or just have value value value 2011-02-12T03:30:22 does a bot send anything other than ant moves? 2011-02-12T03:30:26 no 2011-02-12T03:30:29 "go" 2011-02-12T03:30:36 yes :P 2011-02-12T03:30:41 or better is there any possible future change where it would want to? 2011-02-12T03:30:49 maybe 2011-02-12T03:30:58 so stick with 'O' 2011-02-12T03:31:05 if yes, then stick with a type identifier for bot output too 2011-02-12T03:31:07 let's drop the 'P' 2011-02-12T03:32:10 also, the engine sends 'ready' on turn 0 and 'go' for all others with an 'end' after the game is over 2011-02-12T03:32:38 you could specify that the engine ignores unrecognized types, then people can use it to output debug info mixed in with their orders if they want 2011-02-12T03:32:45 not positive if that's a good idea or not 2011-02-12T03:32:54 no, it's not 2011-02-12T03:33:01 definitely specify that bots should ignore the unrecognized types 2011-02-12T03:33:11 so you can add new types without breaking old bots 2011-02-12T03:33:23 the starter packs should help with that 2011-02-12T03:34:36 be sure to spec it in anyway so no one can cry about it 2011-02-12T03:34:44 even thought they still will... 2011-02-12T03:34:48 though 2011-02-12T03:35:09 janzert: i'm quite happy i get to throw it together just so i can make note of such things 2011-02-12T03:36:45 antimatroid: so, you agree we can drop the 'P' 2011-02-12T03:37:14 wait... 2011-02-12T03:37:23 * janzert knows the last contest made him evil because he'd be tempted to have the engine randomly add invalid types to be sure the bots were correctly rejecting them :P 2011-02-12T03:37:27 I'd kinda like the first char to be a good indicator 2011-02-12T03:37:38 McLeopold: we could drop it 2011-02-12T03:37:50 we can drop the go's but not the ready 2011-02-12T03:37:58 because you ened to give people time to set up etc. 2011-02-12T03:38:12 ened? 2011-02-12T03:38:16 need* 2011-02-12T03:38:28 how could we drop 'go'? 2011-02-12T03:38:36 you know the dimensions 2011-02-12T03:38:36 that's after every turn 2011-02-12T03:38:45 if you read M, get height lines of them 2011-02-12T03:38:56 i'll leave go in, probably makes life easier 2011-02-12T03:38:59 yeah 2011-02-12T03:39:17 because I may only send visible lines 2011-02-12T03:39:31 how do you explain the input in words? 2011-02-12T03:39:55 check what it has now 2011-02-12T03:40:29 ergh, formatting is still hating on me 2011-02-12T03:41:20 each line of type 'M' represents one row of the maps with the (cells/squares) being represented as follows 2011-02-12T03:41:59 something like that anyway is what I would start working from 2011-02-12T03:43:00 I also don't really understand you're radius value definitions 2011-02-12T03:43:57 the value passed is vision^2 2011-02-12T03:44:04 and be sure to send loadtime as the first thing since it is really already in progress when they receive it 2011-02-12T03:44:05 so it can be passed as an int 2011-02-12T03:44:43 say the value passed is the square of the radius then? 2011-02-12T03:44:58 or the actual value sent 2011-02-12T03:45:17 and maybe call the type e.g. viewradius2 2011-02-12T03:45:46 done 2011-02-12T03:45:54 not saved, but it's changed in the edit page 2011-02-12T03:46:06 i'm stripping out sample inputs for the moment 2011-02-12T03:46:19 i can't work out how to do text over multiple lines without paragraph gaps 2011-02-12T03:47:25 oh, now i have i think 2011-02-12T03:47:34 it's markdown right? 2011-02-12T03:47:54 if so leave two trailing spaces on the line 2011-02-12T03:49:05 also need to specify how your location indexing works for ant orders 2011-02-12T03:49:20 e.g. is it zero or one based? 2011-02-12T03:51:03 *** choas has joined #aichallenge 2011-02-12T03:53:20 should we send go after the end E line? 2011-02-12T03:59:28 it might be best to be consistent there 2011-02-12T03:59:54 although I would still just kill the bot when it loses :P 2011-02-12T04:00:43 actually, i think with M, you can just send end info as though it was go 2011-02-12T04:00:47 and process end differently 2011-02-12T04:02:01 probably doesn't make a big difference either way, bots aren't supposed to be doing anything other than shutting down after receiving it anyway 2011-02-12T04:02:26 janzert: the point is to send them information about winners/scores so if people want to train with that information locally, they can 2011-02-12T04:03:03 for training you'll want to get that information to your bot runner though not the bot itself 2011-02-12T04:04:02 but if you want it go for it 2011-02-12T04:06:21 i don't really want it personally 2011-02-12T04:06:35 i figure it's better to give people too much info than not enough 2011-02-12T04:07:23 janzert: can you check the sample input 2011-02-12T04:07:30 does that look good? 2011-02-12T04:07:43 other than the fact that it formatted horribly 2011-02-12T04:08:04 i still need to clean, just trying to get the structure right first 2011-02-12T04:08:49 I still think you should drop P also your example loadtime and turntime look wrong 2011-02-12T04:09:05 ah, they are, so is birth radius for attack 2011-02-12T04:09:56 McLeopold: I am happy to drop P if you are happy for me to? 2011-02-12T04:10:42 if you indent the examples by 4 spaces I think they will be monospaced font so they'll look better 2011-02-12T04:10:51 not positive github does that though 2011-02-12T04:11:13 i committed a their crime 2011-02-12T04:12:11 janzert: that worked a treat thanks 2011-02-12T04:18:19 it's a lot neater now, still needs to be cleaned, but it's readable 2011-02-12T04:22:04 i haven't fixed the examples, they are broken 2011-02-12T04:23:09 are locations using a zero based index? 2011-02-12T04:24:58 I made a little more reasonable anyway 2011-02-12T04:25:05 the example input that is 2011-02-12T04:26:09 yeah locations are zero index based 2011-02-12T04:26:22 i'll put that in at some point, although a programmer shouldn't even need to ask :P 2011-02-12T04:26:45 remember some languages use 1 based indexes for their arrays 2011-02-12T04:27:05 learn* 2011-02-12T04:47:10 antimatroid in your sample input how may i discover from what team is a given ant? 2011-02-12T04:49:41 by the letter given for it 2011-02-12T04:50:05 I believe 'a' is always your own and the other teams start at 'b', 'c', etc. 2011-02-12T04:50:24 ty 2011-02-12T04:50:40 yes, the first enemy you come across will be b for the rest of the game, etc. etc. 2011-02-12T04:50:43 i will put that in after 2011-02-12T04:50:52 what is the significance of a dead ant? 2011-02-12T04:51:04 can i walk into it or eat it or die from it? 2011-02-12T04:51:09 so you can do some scoring internally 2011-02-12T04:51:13 it's equivalent to land 2011-02-12T04:51:16 it's just giving you information 2011-02-12T04:51:58 and if two ants have died in the same spot the information given does not change? 2011-02-12T04:53:13 i don't think you see ants that died from collision 2011-02-12T04:53:20 McLeopold: do you see ants that died from collisions? 2011-02-12T04:53:43 its not important how they died. if its equivilent to land another ant may walk over a dead ant 2011-02-12T04:53:49 and subsequently perish 2011-02-12T04:53:59 pyro-: i expect the attack radius to be large enough that enemy ants can't land on each other 2011-02-12T04:54:08 you can walk onto any dead ant square 2011-02-12T04:54:17 yes 2011-02-12T04:54:22 and some time in the future 2011-02-12T04:54:26 die on the square 2011-02-12T04:54:34 now will i see the old dead ant or the new dead ant? 2011-02-12T04:54:48 if they are from different teams 2011-02-12T04:55:22 pyro-: with a "starter bot" just take a dead ant square to mean it is land 2011-02-12T04:55:24 and leave it at that 2011-02-12T04:55:32 I think you only see a dead ant for the turn after they die, it's basically just to let you know the ant died rather than simple disapeared 2011-02-12T04:55:38 you only see the dead ant the turn after it died 2011-02-12T04:55:43 ah 2011-02-12T04:55:49 ty 2011-02-12T04:55:57 as you don't know where your enemies moved, it'd be hard to work out otherwise 2011-02-12T04:56:07 you don't get information if one of your ants died and it's out of view of ants now 2011-02-12T04:56:12 you need to work that out yourself 2011-02-12T04:56:17 and try to work out who died in the conflict 2011-02-12T04:56:28 clear 2011-02-12T04:56:41 *** Prillicy has quit IRC (Ping timeout: 240 seconds) 2011-02-12T04:56:51 *** Prillicy has joined #aichallenge 2011-02-12T05:00:06 is it safe to ignore the parameter of a P and use the knowledge that the first line is loadtime and the second line is turntime or are the other of parameters subject to change on a game by game basis? 2011-02-12T05:00:17 are the order* 2011-02-12T05:01:01 order should be undefined 2011-02-12T05:01:31 although it's pretty safe that the first one will always be loadtime, still don't assume that 2011-02-12T05:01:56 also the P may go away if I can get that :) 2011-02-12T05:02:08 just then 2011-02-12T05:02:38 there is so much redundancy. P, M, optiontype could all be implicit 2011-02-12T05:03:23 the M is needed 2011-02-12T05:03:33 end 2 0 0 2011-02-12T05:03:34 otherwise it's hard to see if you're getting map data or end of turn info 2011-02-12T05:03:36 there needs to be a type identifier since currently unknown types of lines may be passed in the future 2011-02-12T05:03:38 is not a valid map 2011-02-12T05:03:53 no, the example is broken 2011-02-12T05:03:57 i'll fix that now 2011-02-12T05:04:01 been playing galcon sorry 2011-02-12T05:04:16 janzert: the protocal should be frozen before the competition starts 2011-02-12T05:04:35 that was the problem last time, this time the protocol will be frozen 2011-02-12T05:04:49 it will be all lines are passed as 2011-02-12T05:05:07 you should ignore all types you don't recognize 2011-02-12T05:05:32 if its frozen then im correct in saying that type is redundant :-) 2011-02-12T05:05:47 nope the type of types aren't frozen 2011-02-12T05:05:54 just is frozen 2011-02-12T05:06:22 * janzert is being a bit sarcastic above 2011-02-12T05:06:43 to change the messages that may be sent to a bot after the competition has started is very bad :) 2011-02-12T05:06:48 * janzert really wants to thrown invalid types at bots to make sure the conform now :) 2011-02-12T05:07:02 they 2011-02-12T05:08:33 the message isn't changed as long as you keep ignoring the types you don't know 2011-02-12T05:09:30 on the otherhand... bots that do understand the new types may perform better :p 2011-02-12T05:10:11 sure as long as you can garentee that a bot written with the rules as stated at competition opening are not disadvantaged there is no problem. but if you do so we arrive at the same conclusion that extensibility is not required. 2011-02-12T05:10:57 say there is some problem we discover later on that could be fixed by an extension, what would you do? 2011-02-12T05:11:03 i am going to remove the P's if mcleopold is okay with it 2011-02-12T05:11:16 The turn limit in PlanetWars was such a candidate 2011-02-12T05:11:53 or number of planets 2011-02-12T05:12:20 yeah, too bad if they're disadvantaged but if we're at the point where we need to change we're probably at the point where that is exactly all they get "too bad" and so long as they follow the spec for input at least they'll probably still run 2011-02-12T05:13:15 Frontier: if i have not built the engine and protocol correctly from the start i am an idiot and will cease attempting to organise competitions that i am clearly unable to. i will not modify the rules ad hoc as the competition progresses. 2011-02-12T05:13:48 but to be fair i must say that i would like such changes to happen at least two to four weeks before the deadline 2011-02-12T05:13:58 pyro-: you clearly weren't about at the start of the last contest :P 2011-02-12T05:14:15 what we're doing now was basically done after the contest starte 2011-02-12T05:14:18 wasn't i 2011-02-12T05:14:35 there is always a risk that a problem will crop up with the current game, at least we can try and have the foresight to minimize the impact of any change we have to make 2011-02-12T05:15:46 *** Frontier has quit IRC (Quit: Ex-Chat) 2011-02-12T05:33:13 *** mceier has joined #aichallenge 2011-02-12T05:44:17 *** j3camero has joined #aichallenge 2011-02-12T05:47:32 *** boegel has joined #aichallenge 2011-02-12T05:47:50 i need to say something about distances too 2011-02-12T05:48:18 j3camero: a specification is slowly taking shape on the wiki 2011-02-12T05:48:29 *** boegel has quit IRC (Client Quit) 2011-02-12T05:55:12 i prefer mcleopolds scoring option on kills 2011-02-12T05:55:17 but it's more complicated 2011-02-12T06:29:57 *** Prillicy has quit IRC (Read error: Connection reset by peer) 2011-02-12T06:30:40 *** Prillicy has joined #aichallenge 2011-02-12T06:52:27 i want a different work for food 2011-02-12T06:52:42 *** Prillicy_ has joined #aichallenge 2011-02-12T06:52:45 i have the ants being able to attack each other over water 2011-02-12T06:53:00 but you can't really say they can consume food from across water but not walk over it 2011-02-12T06:53:10 word* 2011-02-12T06:53:39 *** Prillicy has quit IRC (Ping timeout: 240 seconds) 2011-02-12T06:53:50 *** Prillicy_ is now known as Prillicy 2011-02-12T07:05:49 *** mega1 has joined #aichallenge 2011-02-12T07:09:56 *** Prillicy_ has joined #aichallenge 2011-02-12T07:10:16 *** Prillicy has quit IRC (Ping timeout: 276 seconds) 2011-02-12T07:10:22 *** Prillicy_ is now known as Prillicy 2011-02-12T07:15:06 McLeopold: can a "dead" bot from crashing/timing out still accumulate points? 2011-02-12T07:19:32 *** Kingpin13 has joined #aichallenge 2011-02-12T07:27:50 *** Blkt has joined #aichallenge 2011-02-12T07:37:55 good day everyone 2011-02-12T07:41:43 *** Prillicy_ has joined #aichallenge 2011-02-12T07:42:45 *** Prillicy has quit IRC (Ping timeout: 240 seconds) 2011-02-12T07:42:52 *** Prillicy_ is now known as Prillicy 2011-02-12T07:50:14 glorius bacon 2011-02-12T08:30:29 what will i do if an M line is shorter than the specified width 2011-02-12T08:31:51 cry 2011-02-12T08:32:14 assume that will never happen :P 2011-02-12T08:32:27 or you could assume the rest was all unseeable 2011-02-12T08:32:46 if i can assume it will never happen i will just ignore the width paramter 2011-02-12T08:40:40 *** Prillicy_ has joined #aichallenge 2011-02-12T08:41:21 you will need that to make a grid 2011-02-12T08:42:11 i already have the width of the grid from the grid 2011-02-12T08:42:33 sure, but i read chars individually 2011-02-12T08:42:39 so it's easier for me to know the width 2011-02-12T08:42:43 *** Prillicy has quit IRC (Ping timeout: 276 seconds) 2011-02-12T08:42:48 you don't need to make use of every piece of information you're given 2011-02-12T08:42:49 *** Prillicy_ is now known as Prillicy 2011-02-12T08:43:18 of course i won't 2011-02-12T08:43:44 i just always notice the places that i'm making an assumption 2011-02-12T08:44:05 a habbit since thats where i find bugs :) 2011-02-12T08:44:43 if i assume the given width is correct for every M line and it isn't 2011-02-12T08:44:48 things will break :) 2011-02-12T08:45:12 so its not clear what i should do if i detect that 2011-02-12T08:45:22 i will follow your advice and print "cry" :P 2011-02-12T08:52:41 *** Prillicy_ has joined #aichallenge 2011-02-12T08:53:50 *** mceier has quit IRC (Quit: leaving) 2011-02-12T08:54:02 *** Prillicy has quit IRC (Ping timeout: 246 seconds) 2011-02-12T08:54:06 *** Prillicy_ is now known as Prillicy 2011-02-12T08:57:48 :p 2011-02-12T08:58:22 were you pyronaught in pw? 2011-02-12T08:58:26 it claims nz though 2011-02-12T09:01:56 *** delt0r___ has joined #aichallenge 2011-02-12T09:02:36 no 2011-02-12T09:02:46 i gave up on that very early 2011-02-12T09:03:07 but this time i will definitly win!!!!!! ;) 2011-02-12T09:03:10 :p 2011-02-12T09:03:20 *** delt0r_ has quit IRC (Ping timeout: 240 seconds) 2011-02-12T09:04:49 what character(s) indicate end of line under your input protocol? 2011-02-12T09:05:18 hmmm, don't ask me :P 2011-02-12T09:05:22 i don't write the engine 2011-02-12T09:05:29 i would imagine '\n' 2011-02-12T09:05:43 i've never had issues in c++ with input 2011-02-12T09:06:02 have you ever supported unicode? 2011-02-12T09:06:22 me or the contest? 2011-02-12T09:06:25 you 2011-02-12T09:06:34 i don't even really know what unicode is :P 2011-02-12T09:06:47 i'm more a maths/economics student than programmer 2011-02-12T09:06:50 that explains that then 2011-02-12T09:07:07 fair enough 2011-02-12T09:07:29 i better bloody beat you then 2011-02-12T09:07:35 bring it :P 2011-02-12T09:07:44 you're not taking first in australia away from me 2011-02-12T09:09:00 the indignity of being beaten by an economics student would be more than i could stand 2011-02-12T09:09:11 hehe 2011-02-12T09:09:27 a utas student at that :P 2011-02-12T09:09:36 although i'm just doing maths honours this year 2011-02-12T09:13:25 pyro-: the "starter strategy" is going to be as simple as: 2011-02-12T09:13:25 for each ant: try to move north, then east, then south then west 2011-02-12T09:13:38 that way people can download it and make noticable improvements almost straight away 2011-02-12T09:13:51 even beginners to programming 2011-02-12T09:15:23 kk 2011-02-12T09:16:34 im going to sleep now. then i get back from stuff tomorow at early afternoon prolly. then i'll write it. so an ikarus scheme client as per current wiki and above stratagy will exist by this time tomorow 2011-02-12T09:16:38 McLeopold: what happens when there is only one active bot left but ants from more players? do they have to round up all the other ants before they get the food bonus? 2011-02-12T09:17:32 * janzert thinks they should 2011-02-12T09:18:17 if we're also adding to the scores of "dead" bots 2011-02-12T09:18:28 we should probably finish a game where all bots are inactive 2011-02-12T09:18:30 also was going to mention the "food" bonus should include all future food, food currently on the map *and* the players own ants 2011-02-12T09:19:07 is McLeopold zeroing the score of dead ants or just letting it stand where it is? 2011-02-12T09:19:15 stand i think 2011-02-12T09:19:23 although i'm unsure whether the score even freezes 2011-02-12T09:19:39 hmm, I wonder what to do when all the bots have crashed then? 2011-02-12T09:19:53 ^^ that's what i said above, should the game continue running 2011-02-12T09:20:09 it would be fairly fast anyway, as the bots don't need to be given anything, particularly time to think 2011-02-12T09:20:53 seems like a waste of resources to play it out, but it should play out very fast as long as the engine is smart enough not to wait on bots after the first move they timeout on. 2011-02-12T09:21:02 so yeah ;) 2011-02-12T09:21:34 i would be in favour of just zeroing any "crashed" bots score 2011-02-12T09:21:41 the engine should record and give the visualizer some indication on what move a bit crashes on 2011-02-12T09:21:45 then calling the game off and drawing everyone, but i don't really care 2011-02-12T09:22:10 yeah, I think zeroing the score is slightly better but I don't have a strong preference 2011-02-12T09:22:13 I'll add a command for noting the "crash" of a bot 2011-02-12T09:22:36 McLeopold is just making sure he doesn't get hurt as bad as in pw from crashes :P 2011-02-12T09:22:37 jk 2011-02-12T09:22:50 heh 2011-02-12T09:25:59 i think that will do me for tonight with the spec 2011-02-12T09:26:18 the maths doesn't show very nicely for distances 2011-02-12T09:26:33 it wouldn't let me use the align* function with the tex stuff :( 2011-02-12T09:28:59 i think the last unclear thing is battle resolution and then playing around with radius parameters/maps 2011-02-12T09:29:10 and these "end of game" conditions 2011-02-12T09:30:11 yeah, it seems to be basically to the point of needing something running to see how or if things really work out as an interesting game 2011-02-12T09:31:10 even that's unclear till people really get going, but yeah some testing will be good 2011-02-12T09:31:59 *** Accoun has quit IRC (Ping timeout: 276 seconds) 2011-02-12T09:35:45 *** Prillicy_ has joined #aichallenge 2011-02-12T09:37:27 *** Accoun has joined #aichallenge 2011-02-12T09:37:50 *** Prillicy has quit IRC (Ping timeout: 276 seconds) 2011-02-12T09:37:54 *** Prillicy_ is now known as Prillicy 2011-02-12T10:25:19 *** Prillicy has quit IRC (Remote host closed the connection) 2011-02-12T11:16:58 *** amstan has joined #aichallenge 2011-02-12T11:16:58 *** ChanServ sets mode: +o amstan 2011-02-12T11:19:15 amstan: ask and you shall receive 2011-02-12T11:19:20 https://github.com/aichallenge/aichallenge/wiki/Ants-Game-Specification 2011-02-12T11:20:18 antimatroid: decent 2011-02-12T11:20:45 i should probably make an announcement about it 2011-02-12T11:20:47 it's still in need of a clean and revise etc. but it's a start 2011-02-12T11:21:03 we still need engines/visualisers/etc. 2011-02-12T11:21:36 and we still need to convince jeff battle option 1 is superior 2011-02-12T11:22:10 well, make a proposal, have it all written down 2011-02-12T11:22:23 irc is kinda overwhelming if someone is trying to convince you something 2011-02-12T11:22:39 hmm, i might do advantages disadvantages tomorrow 2011-02-12T11:22:42 but i'm going to bed now 2011-02-12T11:22:56 there's also a page for starter bot strategy stuff with possible improvement hints 2011-02-12T12:14:37 antimatroid: you still here? 2011-02-12T12:15:45 so, my thoughts on bot crashes and scores... 2011-02-12T12:16:15 a bot that crashes or times out will stop accumulating points 2011-02-12T12:16:27 the ants will always stay put 2011-02-12T12:16:41 food can still be turned into new ants 2011-02-12T12:17:09 this is too keep the balance of the game, to not give an advantage to one of the other bots still in the game 2011-02-12T12:17:35 the food turning into ants is so that a bot can't tell when another bot has crashed by the game mechanics 2011-02-12T12:18:18 if all but 1 bot crashes, the game ends and the last ant gets the food bonus, plus a bonus for all leftover food and enemy ants 2011-02-12T12:18:42 this would be the same score if we let them clean up the map, so we don't need to drag out the game 2011-02-12T12:19:20 if the last 2 bots crash, the split the bonus, and neither of their ants are included in it 2011-02-12T12:19:38 the score is not zeroed out for a crashed bot 2011-02-12T12:20:24 which means a crashed or timed out bot may still get any ranking in the game 2011-02-12T12:20:53 The reason for this is because I want to emphasize that this is an "AI" contest and not a coding contest. 2011-02-12T12:21:20 Also, there will be no invalid moves, the engine will ignore them. 2011-02-12T12:21:47 We could probably make an engine that would report invalid moves, for testing. 2011-02-12T12:22:06 McLeopold: At a minimum, you probably also want to log timeouts so that people know why the bots didn't do anything. 2011-02-12T12:22:21 yes, timeouts and crashes will be logged 2011-02-12T12:22:35 hopefully on crashes we can capture 1024 bytes of stderr 2011-02-12T12:22:49 McLeopold: Not sure about that one. 2011-02-12T12:23:17 McLeopold: From a security perspective, it's nice for the bots to have no outbound communication channels other than the moves to the engine. 2011-02-12T12:24:01 it would be very useful, there were a lot of forum posts of people confused about why their bot crashed. This would help with that, and hopefully make for less forum maintenance. 2011-02-12T12:24:19 It's not a security issue. 2011-02-12T12:24:48 It's captured as part of the game replay and made available to the owner only. 2011-02-12T12:24:59 McLeopold: Well, it's nice that if bots were to gain access to secret information (other bots' source code, etc) somehow, they would not be able to transmit it. 2011-02-12T12:25:23 Yeah, so, bot owner only. 2011-02-12T12:25:34 McLeopold: I think you're missing my point. 2011-02-12T12:25:37 Plus, bots control what is extracted 2011-02-12T12:25:56 I must be 2011-02-12T12:26:10 If Bot A is able to (by some other security flaw) gain access to Bot B's source code (or our API keys or something), it is nice if we can prevent Bot A from sending that to Bot A's owner. 2011-02-12T12:26:21 The debugging advantage may trump this, though. 2011-02-12T12:26:31 it would 2011-02-12T12:44:06 jbroman: perhaps we should allow bots to read eachother's sources, but block all communications 2011-02-12T12:44:24 jbroman: if they're smart enough to find the best bot and use stuff from its code, so be it, lol 2011-02-12T12:57:14 the visualizer could show "dead" ants differently, maybe 2011-02-12T12:57:24 make their colors more pale maybe 2011-02-12T13:00:59 *** mceier has joined #aichallenge 2011-02-12T13:06:55 jmcarthur: that's what I was thinking 2011-02-12T13:07:07 or just darken the dirt every time an ant dies on a spot 2011-02-12T13:07:19 then the visualizer could show high conflict areas 2011-02-12T13:09:39 heh that'd be neat 2011-02-12T13:09:50 turn it a bit more red, like blood :P 2011-02-12T13:10:24 by "dead" i actually meant when the bot crashes or something 2011-02-12T13:10:29 so the ants are technically still alive 2011-02-12T13:10:36 "frozen" would have been more accurate 2011-02-12T13:12:23 I don't want other bots to be able to determine crashed bots 2011-02-12T13:13:58 What would you think about having warrior ants and gatherer ants? 2011-02-12T13:14:17 It would turn the game into more of an economic challenge 2011-02-12T13:14:24 similar to an rts 2011-02-12T13:15:11 McLeopold: slippery slope 2011-02-12T13:16:09 it's not going to happen, but it would make for a fun game 2011-02-12T13:16:22 you have to balance gather and attack strength 2011-02-12T13:16:48 then we can really call it antz and steal the disney theme 2011-02-12T13:41:32 *** boegel has joined #aichallenge 2011-02-12T14:25:55 i was only talking about the visualizer, not what we send as bot input 2011-02-12T14:27:14 *** sigh has joined #aichallenge 2011-02-12T14:57:36 *** Cyndre_ has quit IRC (Ping timeout: 245 seconds) 2011-02-12T14:59:04 i don't think saving stderr is very feasible in the long run 2011-02-12T14:59:36 *** boegel has quit IRC (Quit: Leaving) 2011-02-12T14:59:48 *** Cyndre has joined #aichallenge 2011-02-12T15:01:55 i'm not sure how i feel about "no invalid moves" 2011-02-12T15:02:25 but we can catch a lot of problems by having a few simple test cases to be sure the bot receives input and gives output 2011-02-12T15:05:30 *** delt0r_ has joined #aichallenge 2011-02-12T15:07:09 *** delt0r___ has quit IRC (Ping timeout: 250 seconds) 2011-02-12T15:07:30 *** flag has joined #aichallenge 2011-02-12T15:14:47 *** Accoun has quit IRC () 2011-02-12T15:17:29 *** delt0r___ has joined #aichallenge 2011-02-12T15:17:44 *** sigh has quit IRC (Remote host closed the connection) 2011-02-12T15:18:43 *** delt0r_ has quit IRC (Ping timeout: 246 seconds) 2011-02-12T15:25:50 *** Accoun has joined #aichallenge 2011-02-12T15:27:37 our x,y coordinates are in quadrant 4. should it be quadrant 1? 2011-02-12T16:11:28 What's the status on the next contest? 2011-02-12T16:24:39 in progress, very close 2011-02-12T16:26:57 Sounds good :) 2011-02-12T16:27:02 Ants right? 2011-02-12T16:27:38 yep 2011-02-12T16:27:47 antimatroid is working on the wiki 2011-02-12T16:27:53 *** j3camero has quit IRC (Read error: Network is unreachable) 2011-02-12T16:28:06 https://github.com/aichallenge/aichallenge/wiki/Ants-Game-Specification 2011-02-12T16:28:20 *** j3camero has joined #aichallenge 2011-02-12T16:28:27 hi j3camero 2011-02-12T16:36:27 *** Guest61834 has joined #aichallenge 2011-02-12T16:36:52 *** Guest61834 has quit IRC (Client Quit) 2011-02-12T17:02:03 *** delt0r_ has joined #aichallenge 2011-02-12T17:03:18 *** delt0r___ has quit IRC (Ping timeout: 250 seconds) 2011-02-12T17:13:34 McLeopold: i like quadrant 1, but i don't think it's that important 2011-02-12T17:14:47 I never worried about it with planet wars 2011-02-12T17:14:56 I just created a distance table and forgot about 2011-02-12T17:14:58 it 2011-02-12T17:15:45 if we go from n,e,s,w to 2 cords for a movement, their might be less confusion? 2011-02-12T17:15:56 or at least people can make consistent mistakes 2011-02-12T17:17:10 honestly, even if you're sending NSEW, internally you are probably using coords anyway 2011-02-12T17:17:16 and for the record, yes, I misspelled 'there' 2011-02-12T17:17:31 and coords, but who cares? ;) 2011-02-12T17:17:48 we will just have to be really clear about the coord system 2011-02-12T17:18:02 0,-1 is north 2011-02-12T17:18:13 yeah, sending coords won't help anyway 2011-02-12T17:18:17 to be more clear i mean 2011-02-12T17:18:33 it would 2011-02-12T17:18:43 if someone got their x and y backwards 2011-02-12T17:18:51 McLeopold: Then it's a trivial fix to their bot. 2011-02-12T17:18:54 you still have to know what coordinates your ants are at, and if it's a map input... 2011-02-12T17:19:06 yes, you're right 2011-02-12T17:19:50 j3camero: ping 2011-02-12T18:11:21 i'm about now.. 2011-02-12T18:12:20 McLeopold: the point of al inactive ants remaining is they may cause food squares to "disappear" and i was curious whether a non-participlating bot can still accumulate points? if not, do all active ants get the whole weight of killing the ants in the future? 2011-02-12T18:23:17 McLeopold: I don't get what you mean about the coordinates being in the fourth quadrant? 2011-02-12T18:23:43 i also changed attack and birth radius to battle and spawn radius 2011-02-12T18:24:03 crashed and timed out bots do not get more points, but they keep the points the earned 2011-02-12T18:24:04 because battle resolution sounds better than attack resolution, then i couldn't use birth 2011-02-12T18:24:31 sounds good 2011-02-12T18:24:35 McLeopold: so what happens to the points of a dying live ant when both active and inactive ants contributed to killing it? 2011-02-12T18:24:54 do the active ants if any exist get to split the winnings between them, or do some of the points go to no one? 2011-02-12T18:25:13 btw i got rid of my second scoring option, i prefer yours anyway 2011-02-12T18:25:17 the points are spread as normal, the inactive ants get the claim some points, but they don't get added to the bot's score 2011-02-12T18:25:41 sure, i'll make the specs reflect that 2011-02-12T18:26:08 so, some of the points go to noone 2011-02-12T18:26:52 the point of inactive bots staying is to keep the game balanced for the others 2011-02-12T18:27:16 McLeopold: i get that, just trying to work out the mechanics of how it will work 2011-02-12T18:27:18 the point of the inactive bot still getting to spawn ants is so that a bot can't detect inactive bots easily 2011-02-12T18:27:21 https://github.com/aichallenge/aichallenge/wiki/Ants-Strategy-Guide 2011-02-12T18:27:34 does that look like a reasonably way for people to be able to get started improving strategy wise? 2011-02-12T18:27:40 it's as if the ants all hold, or the bot is still in the game, but doesn't send order 2011-02-12T18:28:01 no one knows the score until the end 2011-02-12T18:28:59 I think the improvements are too detailed 2011-02-12T18:29:04 But it's a good start 2011-02-12T18:29:15 Here's what I would go with for strategy... 2011-02-12T18:30:00 1) know your coordinate system. (detailed explanation) 2011-02-12T18:30:15 *** Mathnerd314 has joined #aichallenge 2011-02-12T18:30:29 2) the map wraps, know how to deal with this, (same distance and direction algos) 2011-02-12T18:31:18 3) path finding, attempt to calculate the shortest distance between 2 points and find your way around walls 2011-02-12T18:31:56 hmmm yeah okay 2011-02-12T18:31:58 4) balance, find a good mix of ants that will gather food and ants that will attack 2011-02-12T18:32:24 anything talking about specific algos, besides the distance and direction ones, is too detailed 2011-02-12T18:32:24 so focus souly on strategy and let them work out the implementation of it? 2011-02-12T18:32:34 we might point them to common path finding articles though 2011-02-12T18:32:40 yes 2011-02-12T18:32:48 the forums will be filled with implementations 2011-02-12T18:33:04 consider it done 2011-02-12T18:33:08 sweet 2011-02-12T18:33:31 i think the game is pretty much "worked out" other than battle resolution and parameter values 2011-02-12T18:33:44 less unforseen problems 2011-02-12T18:33:52 yeah 2011-02-12T18:34:27 so, if the game is "broken" as it stands, (with option 2) then we either switch to option 1, or do something else 2011-02-12T18:34:42 hmmm, i don't want to have to break option 2 to go with 1 2011-02-12T18:34:59 i am still banking on option 1, surely jeff is willing to at least compromise if everyone else wants 1 2011-02-12T18:35:38 I agree with him that it makes it a much more complicated game. It would be fun, but a little to advanced. 2011-02-12T18:35:59 I think the solution might be to allow full vision and have bots communicate alliances. 2011-02-12T18:36:04 i agree with you borth there, just not about it being too advanced 2011-02-12T18:36:25 i think for either game people are going to be attempting similar things to be near the top, why not pick the more interesting game, even if it's harder 2011-02-12T18:37:36 he's worried about a huge gap in abilities. I think option 1 give a larger spread, but it would still be linear. 2011-02-12T18:37:51 so, I don't think I agree totally with his concerns. 2011-02-12T18:37:56 but I share the vision 2011-02-12T18:38:12 i think he underestimates how hard it will be to "exploit" option 1 too 2011-02-12T18:38:42 He also thinks option 2 doesn't have the possibility of being trivially solved, but I think it might be. 2011-02-12T18:38:50 if anyone uses a tree, i can't see them getting passed using it for any enemy ants at close range 1 turn ahead 2011-02-12T18:39:17 I couldn't see anyone making a useful alpha beta for planet wars, but I was wrong. :) 2011-02-12T18:39:22 i don't think option 2 will be "trivially solved", but there's certainly way less balance making it easier to know what you ought to be doing 2011-02-12T18:39:33 yeah, 2011-02-12T18:39:35 who made a useful alpha beta for pw? 2011-02-12T18:39:41 mega1 2011-02-12T18:39:43 the winner 2011-02-12T18:39:48 only on the first turn 2011-02-12T18:39:56 it worked 2011-02-12T18:40:00 yeah :P 2011-02-12T18:40:12 I thought about maybe splitting the ants into workers and fighters 2011-02-12T18:40:30 that would give you a decision to make on how many you spawn of each 2011-02-12T18:40:53 That would be another change I'd be willing to code if we needed it 2011-02-12T18:40:55 what would the difference be? workers can only be killed not kill? and fighters can't spawn? 2011-02-12T18:41:09 workers can't attack, fighters can't spawn 2011-02-12T18:41:23 so, you need to find the proper mix given your neighbors 2011-02-12T18:41:32 no workers, no growth 2011-02-12T18:41:32 hmmm i like that 2011-02-12T18:41:39 no fighters, no security 2011-02-12T18:41:41 So do i 2011-02-12T18:41:42 but it makes things more complicated 2011-02-12T18:41:43 :) 2011-02-12T18:41:52 so, you would need a spawn command 2011-02-12T18:42:09 you are told you have an "egg" and you choose which to spawn 2011-02-12T18:42:32 then instead of just ant corrds, you get fighter/worker coords 2011-02-12T18:42:48 so, an ant has an owner, now it has an owner and type 2011-02-12T18:43:04 all orders are still the same, except for the spawn order 2011-02-12T18:43:17 can't really display that on a char map? 2011-02-12T18:43:32 lower, upper case 2011-02-12T18:43:42 special symbol for eggs? 2011-02-12T18:43:43 *** mega1 has quit IRC (Ping timeout: 260 seconds) 2011-02-12T18:43:44 that's for dead ants? 2011-02-12T18:44:49 I think it could still be simple enough to understand the rules by watching the game. 2011-02-12T18:46:11 hmmm maybe, i think i'd personally go with with what we have already 2011-02-12T18:46:17 i like just how simple it is 2011-02-12T19:11:46 *** sigh has joined #aichallenge 2011-02-12T19:47:33 *** choas has quit IRC (Quit: leaving) 2011-02-12T20:31:17 *** iFire has quit IRC (Read error: Connection reset by peer) 2011-02-12T20:35:47 TIL parts of america have curfew laws for teenagers 2011-02-12T20:35:48 that's insane 2011-02-12T20:41:56 those curfew laws usually are just for driving 2011-02-12T20:42:14 i've never heard of one that says you must be *at home* by a certain time or anything like that 2011-02-12T20:46:14 the comments on reddit were very much indicative that there are places where teenagers aren't allowed out past liek 11pm 2011-02-12T20:48:21 while i agree some age discrimination isn't necessarily bad (driving, drinking etc.), that is insane 2011-02-12T21:00:05 well, i was unaware of it. that's kind of weird 2011-02-12T21:00:15 my area has a driving curfew for teenagers 2011-02-12T21:00:20 it's like 2am or 3am or something 2011-02-12T21:00:31 i don't know what time the curfew is over though 2011-02-12T21:01:53 ah, no it's midnight 2011-02-12T21:01:59 12am-6am 2011-02-12T21:02:07 if you are under 18 2011-02-12T21:02:11 *** iFire has joined #aichallenge 2011-02-12T21:02:17 unless with a parent 2011-02-12T21:02:36 *** iFire has quit IRC (Read error: Connection reset by peer) 2011-02-12T21:06:37 *** iFire has joined #aichallenge 2011-02-12T21:26:22 jmcarthur: how are people meant to get home from work? 2011-02-12T21:28:37 antimatroid: in victoria you can't get a licence at all until you are 18 2011-02-12T21:28:59 other states it is 17, i think 2011-02-12T21:29:50 *** mceier has quit IRC (Quit: leaving) 2011-02-12T21:29:56 antimatroid: if you're under 18 you're just not working that late 2011-02-12T21:30:34 or a parent has to take you home, but that basically just means you aren't working that late ;) 2011-02-12T21:31:08 jmcarthur: canada has a license requirement for this 2011-02-12T21:31:09 sigh: here it's 16, but in some other states it's 18 2011-02-12T21:31:31 with the learners license, one can't drive between midnight-6am 2011-02-12T21:31:57 *** Blkt has quit IRC (Quit: ERC Version 5.3 (IRC client for Emacs)) 2011-02-12T21:31:58 amstan: i think here in alabama there are laws about how late minors can work and stuff, too, even apart from driving 2011-02-12T21:32:05 here learners requires a fully-licenced supervisor, is it different over there? 2011-02-12T21:32:12 sigh: that too 2011-02-12T21:32:54 *** Blkt has joined #aichallenge 2011-02-12T21:33:27 sigh: yeah learners requires at least 15 years of age and a parent or licensed driver over 18, and when you turn 16 if you don't upgrade to a license you can drive with any supervisor with a license 2011-02-12T21:33:41 i think i mixed something up there 2011-02-12T21:33:51 but you get the gist 2011-02-12T21:34:16 yeah, similar here 2011-02-12T21:35:01 supervisor just needs to have a full licence (and not be over blood alcohol limit :P) 2011-02-12T21:36:16 why are all the radii named radius2? is there a viewradius1? 2011-02-12T21:37:45 oh that's sneaky... it's like that prank where you release pigs numbered 1,2 and 4 2011-02-12T21:37:58 * sigh makes a note in case I want to troll a code base sometime 2011-02-12T21:39:28 pyro-: maybe it's meant to be "radius squared" 2011-02-12T21:40:13 oh right 2011-02-12T21:40:15 makes sense 2011-02-12T21:40:18 thx 2011-02-12T21:40:45 * jmcarthur is tempted to make a bot in Agda, for kicks 2011-02-12T21:41:03 *** Blkt` has joined #aichallenge 2011-02-12T21:41:18 the same naming scheme isn't used in the visualiser input 2011-02-12T21:41:21 sigh: i didn't know p's was 18 in victoria 2011-02-12T21:41:40 my p's just ran out yesterday, i need to go renew my license then i'll be fully licensed finally 2011-02-12T21:41:46 i was lazy with getting my license 2011-02-12T21:41:46 antimatroid: good afternoon 2011-02-12T21:41:47 i thought of that because agda allows unicode characters and allows arbitrary characters to be used in identifiers (only whitespace matters for seperating identifiers) 2011-02-12T21:42:07 so you could have r² for radius squared 2011-02-12T21:42:18 *** Blkt` has quit IRC (Remote host closed the connection) 2011-02-12T21:42:19 *** Blkt has quit IRC (Remote host closed the connection) 2011-02-12T21:42:33 pyro-: it is the radius's squared 2011-02-12T21:42:43 i 2011-02-12T21:42:50 i'll fix the visualiser stuff in a bit 2011-02-12T21:44:25 McLeopold: we need to decide whether blockages are walls or water 2011-02-12T21:44:32 i think water makes more sense with fighting going over it 2011-02-12T21:45:12 if anyone has a strong preference I'm fairly indifferent, so I'm happy to go with what other people want 2011-02-12T21:45:22 why water? 2011-02-12T21:45:40 dunno, fighting and spawning aren't blocked by it 2011-02-12T21:45:42 but walking is 2011-02-12T21:45:54 what's an intuitively reasonable thing to call them? 2011-02-12T21:46:29 apparently the game making sense to a real world situation is important 2011-02-12T21:46:29 tall grass? I dunno, somehow I can't really see ants being surrounded by water 2011-02-12T21:46:45 it rains in a forest doesn't it? 2011-02-12T21:47:01 i suppose 2011-02-12T21:47:02 i imagine there are times where their movement is contrained by small water regions 2011-02-12T21:47:24 but it doesn't really make much more sense for them to fight/spawn over those regions than to walk 2011-02-12T21:47:29 Why not have the ants going through tunnels and the rest be the walls? 2011-02-12T21:47:38 what? 2011-02-12T21:47:59 As in, the moveable area is a tunnel, the non-moveable/spawnable area is the wall of the tunnel 2011-02-12T21:48:04 such as in an anthill 2011-02-12T21:49:12 the point is people can fight an ant that is on the other side of a wall 2011-02-12T21:49:25 is that a requirement? 2011-02-12T21:49:31 that's how the game currently works 2011-02-12T21:49:39 and i think it'd be easier to block regions off otherwise 2011-02-12T21:49:48 which people seem to think is unideal 2011-02-12T21:49:48 quicksand then maybe? 2011-02-12T21:50:30 maybe if spawning was blocked by walls we could argue the walls aren't so high they can't fight 2011-02-12T21:50:58 in fact, it would probably be better if spawning was blocked by walls 2011-02-12T21:51:20 or if spawn radius < 2 then you can't get a food square in range through a wall 2011-02-12T21:51:35 i think that's ideal for spawn radius anyway 2011-02-12T22:06:37 McLeopold: give me a yell when you're about, i want to talk food bonus conditions again 2011-02-12T22:39:12 we could get beginners to try and gang up on the best player 2011-02-12T22:39:57 i can't decide if i want to represent the map as a group of linked cells or an array 2011-02-12T22:40:59 how will you know who the best player is? 2011-02-12T22:41:08 go for the player with the most ants you can see 2011-02-12T22:41:11 and only them 2011-02-12T22:41:21 you can't see shit about who has how many ants with the fog 2011-02-12T22:41:21 or most ants you've seen 2011-02-12T22:41:40 you can with the radius being sqrt(96) and a few ants 2011-02-12T22:46:13 i think i'll go with an array 2011-02-12T22:46:59 oh man, i almost want to not do an actual bot and make my bot the ultimate troll 2011-02-12T22:47:29 i'm at least making one for tcp 2011-02-12T22:56:10 *** bhasker has joined #aichallenge 2011-02-12T23:01:04 *** jmcarthur has quit IRC (Quit: WeeChat 0.3.2) 2011-02-12T23:12:10 *** jmcarthur has joined #aichallenge 2011-02-12T23:29:25 *** j3camero has quit IRC (Quit: leaving) 2011-02-12T23:44:04 *** Prillicy has joined #aichallenge 2011-02-12T23:44:39 finally added a signature 2011-02-12T23:51:06 i guess i should yet again change how my bot takes input