2011-06-05T00:22:49 *** filter has quit IRC (Ping timeout: 240 seconds) 2011-06-05T00:41:28 *** filter has joined #aichallenge 2011-06-05T00:42:56 *** delt0r_ has joined #aichallenge 2011-06-05T00:44:11 *** delt0r___ has quit IRC (Ping timeout: 240 seconds) 2011-06-05T01:14:44 *** JaycobColeman has joined #aichallenge 2011-06-05T01:18:54 Back from Canada! Any monumental changes happen in the last couple weeks? 2011-06-05T01:25:25 JaycobColeman: yes, we're almost ready! 2011-06-05T01:26:06 That's wonderful! Also, you live in Canada yes? 2011-06-05T01:27:35 yes 2011-06-05T01:27:38 hamilton 2011-06-05T01:27:42 jeff lives in toronto 2011-06-05T01:30:57 I see. Unfortunately I never made it so far east :/ 2011-06-05T01:31:11 mostly just spent time in the yukon and BC 2011-06-05T01:31:15 a little in alberta 2011-06-05T01:31:42 Wish I could've explored more, but I was really just passing through to washington 2011-06-05T01:34:04 *** Meatkat__ has joined #aichallenge 2011-06-05T01:37:12 *** Meatkat_ has quit IRC (Ping timeout: 244 seconds) 2011-06-05T01:44:53 Interesting fact: Every Canadian I met on my way down and on the way back was incredibly nice. ALL OF THEM 2011-06-05T01:45:24 Screw America, I'm moving to Canada 2011-06-05T01:59:01 lol 2011-06-05T01:59:31 i found a cool graphic artist that draws ants 2011-06-05T01:59:44 i was thinking we could add some ant artwork in our site 2011-06-05T01:59:48 example: http://home.hypertriangle.com/~alex/aichallenge/website/ 2011-06-05T01:59:51 [11:44] xQuasar: and you call yourself a programmer1 <-- Referring to what? the array indices? 2011-06-05T02:00:09 xQuasar: yes, most programming languages start from 0 without a question 2011-06-05T02:00:23 I was just making sure xD 2011-06-05T02:01:45 Also I don't even call myself a programmer, i'm a high school student 2011-06-05T02:01:47 :p 2011-06-05T02:06:20 *** JaycobColeman has quit IRC (Ping timeout: 252 seconds) 2011-06-05T02:08:05 xQuasar: he was just teasing you, don't worry 2011-06-05T02:08:56 amstan: I know, I was just sayin' :P 2011-06-05T02:11:14 contestbot: later tell mcleopold see https://github.com/aichallenge/aichallenge/issues/162 2011-06-05T02:11:14 amstan: As you wish. 2011-06-05T02:11:18 contestbot: later tell janzert see https://github.com/aichallenge/aichallenge/issues/162 2011-06-05T02:11:18 amstan: Ready to serve. 2011-06-05T02:11:23 anyone else, see that ^ 2011-06-05T02:27:28 *** xQuasar has quit IRC (Ping timeout: 240 seconds) 2011-06-05T02:32:09 *** xQuasar has joined #aichallenge 2011-06-05T02:45:27 *** xQuasar has quit IRC (Ping timeout: 252 seconds) 2011-06-05T02:50:06 *** xQuasar has joined #aichallenge 2011-06-05T02:59:04 *** Palmik has joined #aichallenge 2011-06-05T03:01:04 *** xQuasar has quit IRC (Ping timeout: 252 seconds) 2011-06-05T03:02:30 *** UncleVasya has joined #aichallenge 2011-06-05T03:06:22 *** xQuasar has joined #aichallenge 2011-06-05T03:11:50 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-06-05T03:20:09 *** Quasar has joined #aichallenge 2011-06-05T03:20:35 *** Quasar is now known as Guest70020 2011-06-05T03:22:29 *** Meatkat_ has joined #aichallenge 2011-06-05T03:22:48 *** xQuasar has quit IRC (Ping timeout: 260 seconds) 2011-06-05T03:25:38 *** Meatkat__ has quit IRC (Ping timeout: 252 seconds) 2011-06-05T03:53:42 *** JaycobColeman has joined #aichallenge 2011-06-05T04:15:32 *** JaycobColeman has quit IRC (Quit: Page closed) 2011-06-05T04:32:02 *** aerique has joined #aichallenge 2011-06-05T04:45:44 *** needsch has joined #aichallenge 2011-06-05T04:47:23 *** qacek has quit IRC (Quit: WeeChat 0.3.5) 2011-06-05T05:03:22 *** needsch1 has joined #aichallenge 2011-06-05T05:06:39 *** needsch has quit IRC (Ping timeout: 260 seconds) 2011-06-05T05:11:52 *** Guest70020 has quit IRC (Ping timeout: 260 seconds) 2011-06-05T05:12:12 *** xQuasar has joined #aichallenge 2011-06-05T05:16:16 *** Accoun has quit IRC (Ping timeout: 244 seconds) 2011-06-05T05:17:20 *** xQuasar has quit IRC (Ping timeout: 276 seconds) 2011-06-05T05:17:28 *** Accoun has joined #aichallenge 2011-06-05T05:20:08 *** xQuasar has joined #aichallenge 2011-06-05T05:41:03 *** FireFly has joined #aichallenge 2011-06-05T06:03:27 *** yasith has joined #aichallenge 2011-06-05T06:08:11 *** aerique has quit IRC (Ping timeout: 240 seconds) 2011-06-05T06:15:18 *** Redgis has joined #aichallenge 2011-06-05T06:30:48 *** Frontier has joined #aichallenge 2011-06-05T06:36:20 *** Frontier has quit IRC (Ping timeout: 240 seconds) 2011-06-05T06:39:53 *** Frontier has joined #aichallenge 2011-06-05T07:25:41 *** needsch1 has quit IRC (Ping timeout: 240 seconds) 2011-06-05T07:30:42 *** needsch has joined #aichallenge 2011-06-05T07:49:31 *** smiley1983 has joined #aichallenge 2011-06-05T07:49:56 antimatroid: you asked if I had looked at fighting yet 2011-06-05T07:50:11 i ask anyone that :P 2011-06-05T07:50:12 but go on... 2011-06-05T07:50:56 A little, but my current thoughts are unimplemented, pending completion / extension of certain functions and data structures 2011-06-05T07:51:27 yeah, i haven't implemented anything fighting wise, but i'm still curious about thoughts on it 2011-06-05T07:51:49 once my exams are over and i've fixed up maps i'll try to churn out a semi-decent bot and then call it a day with ants 2011-06-05T07:51:52 I've done some of the obvious searching for bad moves, good moves, and adjusting accordingly 2011-06-05T07:51:53 i don't have a lot of time this year 2011-06-05T07:52:02 how are you doing that? 2011-06-05T07:52:15 are you checking for combinations of their ants moves? moving their ants first? or just assuming their ants don't move? 2011-06-05T07:52:48 checking a few states generated based on different assumptions 2011-06-05T07:53:06 one assumption is that the enemy doesn't move 2011-06-05T07:53:33 one assumption is that the enemy tries to fill attack squares as aggressively as possible 2011-06-05T07:53:37 how many ants are you checking moves for at a time? just one? if so how do you order the ants for making move selection? closest/furthest enemy? 2011-06-05T07:53:40 but only looks at one such state 2011-06-05T07:53:45 if you are doing multiple, are you doing them all or partitioning? 2011-06-05T07:54:04 yeah, i think i'm pretty much planning those two assumptions 2011-06-05T07:54:12 then maximise my minimum outcome 2011-06-05T07:54:33 Whole map in one pass, but my "proposed orders" are cancelled recursively if they have bad outcomes 2011-06-05T07:54:56 so that iteration has a sort of partitioning effect 2011-06-05T07:55:22 have you thought much about macro strategy? ie. spreading out, maximising visibility/fastest access to squares, pairing up friend ants to enemy ants etc? 2011-06-05T07:55:29 yes 2011-06-05T07:55:33 a bit 2011-06-05T07:56:11 all unimplemented 2011-06-05T07:56:28 yeah i haven't implemented any of this stuff, i'm still just thinking about ideas 2011-06-05T07:56:52 it would be awesome to make a bot that partitions "fight" ants and does full turn game trees, but we don't have enough time with the number of ants :( 2011-06-05T07:57:01 I've been interested to see some different techniques in action on the beta server 2011-06-05T07:57:01 it would be sweet to watch some games with such a bot though 2011-06-05T07:57:23 are there some? i haven't watched any games in like a week or two 2011-06-05T07:58:02 I'm not sure it would gain a huge benefit from it - some, but a better expansion system will still probably beat it without deep combat analysis 2011-06-05T07:58:24 i'm pretty happy with my food collection 2011-06-05T07:58:30 some expansion strategies can be seen 2011-06-05T07:58:34 a few minor tweaks needed, but otherwise it's pretty solid 2011-06-05T07:58:46 care to offer any thoughts on food collection? 2011-06-05T07:59:04 i'll explain my basic food collection algorithm 2011-06-05T07:59:17 :) 2011-06-05T08:00:54 create an "available" map for the grid, setting current ant locations at 0 and non-ant locations at unavailable, then from the set of not yet collected food, iteratively do a bfs from all food locations at once to find the closest ant that can make it within the spawnradius of a food square, "move the ant", update the availability information at it's current square, target square and the food square, then collecting the 2011-06-05T08:01:13 you can decide whether you want to collect food with enemy ants or not (this is a strategy preference) 2011-06-05T08:02:04 you need to keep the bfs going until like currentDist >= bestDist + bestLocationAvailability 2011-06-05T08:02:47 it's not optimal, but i don't see people getting much closer to optimal without a tree like structure 2011-06-05T08:03:01 do you repeat the bfs after updating availability info? 2011-06-05T08:03:05 yes 2011-06-05T08:03:16 ah, I hadn't thought of that 2011-06-05T08:03:36 it can take upwards of 200ms with lots of food and ants, but you can "call it off after x amount of time" if you want, and you will still have collected the closest food 2011-06-05T08:03:47 yes 2011-06-05T08:04:32 it collects food like 5-6 times faster than hunterbot 2011-06-05T08:04:42 if there's lots of it 2011-06-05T08:04:55 the one I have on the server is very silly 2011-06-05T08:05:27 it sends every ant after its closest food, even if they all go after one 2011-06-05T08:06:26 :P, you could at least stop going after each piece of food when you've gone for it :P 2011-06-05T08:06:30 but mine even improves on that 2011-06-05T08:07:25 No, it's just doing a bfs from food items, marking the distance, and the every ant steps to the tile with the smallest food distance 2011-06-05T08:07:53 No point improving it, though, that code is obsolete :) 2011-06-05T08:09:34 one question about mine is whether to use enemy ants while collecting food 2011-06-05T08:09:57 assuming the enemy employs a similar strategy, you probably want to, otherwise it may work better to go after food items nearer to them first 2011-06-05T08:11:53 it will be interesting to see how the food collection strategies develop wrt enemies 2011-06-05T08:12:08 do they have a planned duration? 2011-06-05T08:12:28 i think people have been pretty happy with my suggestion of 2 months 2011-06-05T08:12:41 tron was too short and pw too long from the general consensus i think 2011-06-05T08:12:49 so i've applied the goldy locks rules 2011-06-05T08:12:58 haha 2011-06-05T08:13:21 *** yasith has quit IRC (Remote host closed the connection) 2011-06-05T08:13:36 yes, I think 2 months sounds good 2011-06-05T08:13:47 *** yasith has joined #aichallenge 2011-06-05T08:14:27 i want to get started on zeta after that 2011-06-05T08:14:46 as in multi-contest framework? 2011-06-05T08:14:51 i think after ants we should say no more contests until zeta is up and running, then hopefully get servers up for old games and a few others 2011-06-05T08:14:58 then once that is somewhat done do another contest 2011-06-05T08:14:59 yeah 2011-06-05T08:16:39 yes, I would love to develop a tron bot, and if pw were back I would return to it eventually... plus I know others who missed that contest and would like to develop a pw bot 2011-06-05T08:17:36 yeah, i'm super keen to play around with tron again 2011-06-05T08:17:43 you actually get to use a proper game tree 2011-06-05T08:18:20 I've never actually implemented one before, so I'd like to 2011-06-05T08:18:24 one of the harder problems with tron is longest path once you separate into regions 2011-06-05T08:18:43 if we did it again, we'd have to do it multiplayer with some 2-player maps 2011-06-05T08:19:33 and chambers can be tricky, but not too hard to do somewhat well 2011-06-05T08:19:42 would it be hard to make that fair? (multiplayer tron) 2011-06-05T08:19:48 that being, if you leave a chamber too early, you can't get back in to fill the space up 2011-06-05T08:19:56 no, just do symmetric maps 2011-06-05T08:20:08 i could use the ants maps i think 2011-06-05T08:20:25 you'd want more open maps though 2011-06-05T08:20:34 as the walls are created as the game progresses 2011-06-05T08:22:07 I hope to see zeta running then. :) (responses becoming more intermittent, eating) 2011-06-05T08:25:42 *** aeter has joined #aichallenge 2011-06-05T08:29:23 *** xQuasar has quit IRC (Ping timeout: 252 seconds) 2011-06-05T08:31:53 *** CIA-48 has quit IRC (Ping timeout: 276 seconds) 2011-06-05T08:35:08 *** Meatkat__ has joined #aichallenge 2011-06-05T08:35:28 *** xQuasar has joined #aichallenge 2011-06-05T08:36:14 *** CIA-26 has joined #aichallenge 2011-06-05T08:38:21 *** Meatkat_ has quit IRC (Ping timeout: 260 seconds) 2011-06-05T08:43:24 *** delt0r___ has joined #aichallenge 2011-06-05T08:45:03 *** delt0r_ has quit IRC (Ping timeout: 255 seconds) 2011-06-05T08:50:11 *** Quasar has joined #aichallenge 2011-06-05T08:50:37 *** Quasar is now known as Guest26797 2011-06-05T08:53:36 *** xQuasar has quit IRC (Ping timeout: 255 seconds) 2011-06-05T08:57:46 *** yasith has quit IRC (Read error: Operation timed out) 2011-06-05T09:04:57 *** Accoun has quit IRC (Ping timeout: 240 seconds) 2011-06-05T09:11:57 *** yasith has joined #aichallenge 2011-06-05T09:12:51 *** olexs has joined #aichallenge 2011-06-05T09:13:27 *** sigh has joined #aichallenge 2011-06-05T09:19:04 *** Accoun has joined #aichallenge 2011-06-05T09:19:51 *** Eruonen has joined #aichallenge 2011-06-05T09:24:29 *** UncleVasya has quit IRC (Ping timeout: 260 seconds) 2011-06-05T09:25:40 *** Accoun has quit IRC (Ping timeout: 240 seconds) 2011-06-05T09:27:07 *** Accoun has joined #aichallenge 2011-06-05T09:31:20 *** Accoun has quit IRC (Ping timeout: 250 seconds) 2011-06-05T09:31:55 *** locutus2 has joined #aichallenge 2011-06-05T09:35:34 *** Accoun has joined #aichallenge 2011-06-05T09:54:31 *** UncleVasya has joined #aichallenge 2011-06-05T10:11:41 * antimatroid is cold, accidentally turns heater onto fan setting, damnit! 2011-06-05T10:16:24 *** berak has joined #aichallenge 2011-06-05T10:16:26 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-06-05T10:24:05 *** sigh has quit IRC (Remote host closed the connection) 2011-06-05T10:32:09 *** berak has quit IRC (Ping timeout: 260 seconds) 2011-06-05T10:49:25 *** Meatkat__ has quit IRC (Ping timeout: 240 seconds) 2011-06-05T10:55:55 *** Bhaskerh has joined #aichallenge 2011-06-05T10:57:07 Morning 2011-06-05T11:01:33 morning 2011-06-05T11:01:50 i have to pick an exam up in like 9 hours : 2011-06-05T11:01:52 :\* 2011-06-05T11:02:07 it's weird studying for something i don't need to memorise anything for 2011-06-05T11:02:54 *** Meatkat__ has joined #aichallenge 2011-06-05T11:06:06 All the best:) 2011-06-05T11:07:15 I bought order and chaos online for the ipad 2011-06-05T11:07:32 Nice wow clone 2011-06-05T11:14:12 :P 2011-06-05T11:14:14 i've never played wow 2011-06-05T11:14:24 how's akshara going? 2011-06-05T11:14:28 is she in a bit of a routine now? 2011-06-05T11:14:58 reminds me of akbar (forgive me :P) 2011-06-05T11:32:55 Akshara is doing great , sleeps better at night but not for very long stretches 2011-06-05T11:37:45 nice 2011-06-05T11:45:28 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-06-05T11:48:36 *** ohcrivens has joined #aichallenge 2011-06-05T11:52:04 *** Bhaskerh has quit IRC (Quit: AndroIRC) 2011-06-05T11:52:31 *** Bhaskerh has joined #aichallenge 2011-06-05T12:09:05 *** Meatkat__ has quit IRC (Ping timeout: 244 seconds) 2011-06-05T12:13:26 *** berak has joined #aichallenge 2011-06-05T12:18:23 *** Frontier has quit IRC (Ping timeout: 246 seconds) 2011-06-05T12:21:59 *** aeter has left #aichallenge 2011-06-05T12:23:04 *** needsch has quit IRC (Quit: Leaving.) 2011-06-05T12:32:36 *** ohcrivens has quit IRC (Ping timeout: 252 seconds) 2011-06-05T12:38:24 *** smiley1983 has quit IRC (Quit: Lost terminal) 2011-06-05T12:38:36 heh, i just realized that frozen synapse is very much like a game idea that i've been wanting to have a contest around 2011-06-05T12:41:28 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-06-05T12:42:48 *** _0x471 <_0x471!~slash@p57B3FB61.dip.t-dialin.net> has joined #aichallenge 2011-06-05T12:44:56 *** _0x47 <_0x47!~slash@p5B12E654.dip.t-dialin.net> has quit IRC (Ping timeout: 258 seconds) 2011-06-05T12:54:46 *** Naktibalda has joined #aichallenge 2011-06-05T13:08:20 *** berak has quit IRC (Read error: Connection reset by peer) 2011-06-05T13:10:02 jmcarthur: yea--its basically the squad idea more or less 2011-06-05T13:14:29 ? 2011-06-05T13:17:14 *** Redgis has quit IRC (Quit: ... mains libres) 2011-06-05T13:23:24 *** ohCrivens has joined #aichallenge 2011-06-05T13:26:46 *** mcstar has joined #aichallenge 2011-06-05T13:26:54 antimatroid: it was a good while ago. basically it's a squad combat game with continuous time and continuous space where you issue all the orders for some duration of time, and then the game is advanced that amount of time. it keeps the simultaneous, turn-based play that we normally have in this contest but allows for continuous time and space games 2011-06-05T13:27:42 ah yeah, would like 2011-06-05T13:27:49 there's many types of ai games i would like to explore 2011-06-05T13:28:06 hopefully with zeta we can put up some more advanced games that aren't actually run as an official contest 2011-06-05T13:29:33 my biggest fear about that is an overall lower number of bots submitted for each game, even officially supported games, due to people choosing other games 2011-06-05T13:30:01 yep 2011-06-05T13:30:09 i want it to be more of a training/learning center 2011-06-05T13:30:13 i'm not completely against the idea. i just don't know how badly my fears will come true 2011-06-05T13:30:27 i don't think it would be too hard to spark interest with an official contest game once a year or so 2011-06-05T13:30:58 it would also allow contest games to possibly be a bit more complicated 2011-06-05T13:31:06 yeah i think if we make it rare enough and loud enough then the official games should get a lot of attention 2011-06-05T13:31:09 if people have servers they can play around with say tron and pw with first 2011-06-05T13:31:13 build their way up 2011-06-05T13:31:39 tron is the best starting ai game imo 2011-06-05T13:31:48 (sponsors sponsors and recruitment) 2011-06-05T13:31:48 biased because it was mine, but it was so awesome 2011-06-05T13:32:07 i lubbed tron 2011-06-05T13:32:16 lubed :P 2011-06-05T13:32:18 ew 2011-06-05T13:37:18 antimatroid: I think PW is easier to start than Tron because it is more intuitive. PW can provide more ideas to implement for the bot before things turn into the mathematics. 2011-06-05T13:37:36 erm? 2011-06-05T13:37:39 did you compete in tron? :P 2011-06-05T13:37:43 i don't even see that argument 2011-06-05T13:37:57 tron was fairly easy game tree search 2011-06-05T13:38:15 with longest path once you separate, also doable with a decision tree i suspect 2011-06-05T13:38:22 planet wars was a beast of a problem 2011-06-05T13:38:26 tron was a simpler game, required less math, was easier to visualize, was easier to improve upon, didn't require resorting to so much guesswork 2011-06-05T13:38:52 heck, each turn aside from the first only gave you three possible choices 2011-06-05T13:38:55 *up to 2011-06-05T13:39:07 i'm with jmcarthur :P 2011-06-05T13:39:32 *** olexs1 has joined #aichallenge 2011-06-05T13:41:00 i'm riding the karma train today 2011-06-05T13:41:17 *** olexs has quit IRC (Ping timeout: 276 seconds) 2011-06-05T13:44:16 I mean easier not to be best, but to get started. You say word "tree". From the very beginnig of the development bot for Tron contestant has to think about it. While in PW before contestant will need it he can implement some intuitive scripting things like "go to front", "snipe neutrals", "defend planets" etc. When he realize that scripting sucks he will start thinking about building some trees. 2011-06-05T13:45:02 But maybe I didn't understand what you mean by "best starting ai-game". 2011-06-05T13:49:23 *** needsch has joined #aichallenge 2011-06-05T13:49:32 you don't need a game tree just to get started with tron 2011-06-05T13:49:48 all you need for tron is up, down, left, or right 2011-06-05T13:50:27 i think what you mean is that PW has more stuff to obscure the tree, which IMO makes it harder rather than easier 2011-06-05T13:52:08 *** Bhaskerh has quit IRC (Ping timeout: 276 seconds) 2011-06-05T13:52:32 "best starting ai game" to me means that the game is easy to learn, easy to improve with, hard to master, and encourages you to survey a lot of prior research into similar games while still encouraging creativity on your own 2011-06-05T13:53:07 tron was all of that, to me, but PW lacked all that except for being hard to master and encouraging creativity 2011-06-05T13:54:41 my prediction for ants is that it's somewhere between tron and PW as far as ease of learning, harder to improve upon and master than both, between them on research, and about on par with PW on creativity 2011-06-05T13:55:42 huh, that makes it sound like i like ants more than PW, but that's not really the case... 2011-06-05T13:55:54 maybe i need to add something else to my criteria 2011-06-05T13:56:46 *** Meatkat has joined #aichallenge 2011-06-05T13:58:18 eh, i think the "hard to improve upon" aspect is the main thing making be really not like ants 2011-06-05T13:58:28 *making me 2011-06-05T14:05:54 *** Meatkat has quit IRC (Ping timeout: 255 seconds) 2011-06-05T14:09:30 *** nann has joined #aichallenge 2011-06-05T14:10:59 hi, the highwater mark entry for a bot in the ranking seams not always correct 2011-06-05T14:12:36 for my bot 'locutus2' it should be 43.58 instead 35.17 which is currently displayed 2011-06-05T14:13:26 *** Apophis_ has quit IRC (Ping timeout: 250 seconds) 2011-06-05T14:15:33 i think pw was a cooler "problem" to ants 2011-06-05T14:15:39 but i think ants will work better as an ai game 2011-06-05T14:19:15 locutus2: I think it's the last score of your previous version, not the best one. 2011-06-05T14:24:50 no, if you check bhasker2 you have 44.85 vs. 38.43 2011-06-05T14:25:39 i checked some others bots too and there was it already the all time high 2011-06-05T14:34:21 *** Accoun has quit IRC () 2011-06-05T14:35:46 *** amstan_ has joined #aichallenge 2011-06-05T14:35:46 *** ChanServ sets mode: +o amstan_ 2011-06-05T14:39:02 morning 2011-06-05T14:40:20 i think ants has less strategy than planetwars frankly 2011-06-05T14:40:35 yeah 2011-06-05T14:40:54 my logic with ants is already at a point where i doubt i will improve it much further 2011-06-05T14:40:59 combat is too simplistic 2011-06-05T14:41:09 unless someone can really work out formations very well 2011-06-05T14:41:13 i doubt most bots will get better 2011-06-05T14:42:01 and mostly the problem seems to be due to one guy growing exponentially 2011-06-05T14:42:07 if you are lucky and get a weak neighbour 2011-06-05T14:42:09 that bot is bound to win 2011-06-05T14:42:11 pretty much 2011-06-05T14:42:17 as long as it doesn't do something stupid 2011-06-05T14:42:41 maybe restricting ants to a max of 4 players would be better 2011-06-05T14:42:45 7-10 player games are too random 2011-06-05T14:42:56 and i really doubt they add more value to the skill rating 2011-06-05T14:43:28 a mediocre bot could easily bot a superior bot if it got enough growth and territory control 2011-06-05T14:43:39 due to its neighbours being crappy 2011-06-05T14:43:46 easily beat* 2011-06-05T14:46:25 also planetwars had some really interesting bots to watch, like mcleopolds ant bot/scotbot, etc and the formations were cool 2011-06-05T14:46:30 ants somehow seems to lame to watch 2011-06-05T14:46:31 *** Accoun has joined #aichallenge 2011-06-05T14:46:57 what would also be interesting is to see if increasing the # of starter ants makes a difference 2011-06-05T14:47:06 how do bots do if given 5-6 ants at start instead of just 1 2011-06-05T14:47:26 *** olexs has joined #aichallenge 2011-06-05T14:48:05 morning olexs 2011-06-05T14:48:53 *** olexs1 has quit IRC (Ping timeout: 276 seconds) 2011-06-05T14:50:34 evening bhasker :) 2011-06-05T14:50:40 :) 2011-06-05T14:50:44 eu? 2011-06-05T14:50:46 or east coast? 2011-06-05T14:50:52 eu 2011-06-05T14:50:56 ah 2011-06-05T14:52:06 ah damn just when i upload a new version the worker stops 2011-06-05T14:52:07 ! 2011-06-05T14:52:10 *** amstan_ has quit IRC (Ping timeout: 260 seconds) 2011-06-05T14:53:05 ah started again 2011-06-05T14:53:31 compiling? 2011-06-05T14:53:43 no idea the server statistics page showed 0 games last 5 minutes 2011-06-05T14:53:46 and didn't show an active server 2011-06-05T14:53:48 now it shows one 2011-06-05T14:53:56 http://aichallengebeta.hypertriangle.com/server_stats.php 2011-06-05T14:53:57 workers do both game playing and compiling, afaik 2011-06-05T14:54:16 the top #1 is my ghost bot 2011-06-05T14:54:21 maybe worker that gets a compile job stops showing on statistics because it can't play games 2011-06-05T14:54:22 from like 3 days back 2011-06-05T14:54:31 *** Jaggu has joined #aichallenge 2011-06-05T14:54:39 weird that trueskill levels are not climbing anymore 2011-06-05T14:54:49 seem to be stuck in the low 40's high 30's from the top bots 2011-06-05T14:54:53 they are sinking the whole time, something isn't quite right imho 2011-06-05T14:55:20 before my bot got any good, sir_macelon was >50 2011-06-05T14:55:34 heh yea 2011-06-05T14:55:48 but it could be that most top bots are beating each other on different maps 2011-06-05T14:55:51 but I like how it's c 2011-06-05T14:55:57 c++ on the top 2011-06-05T14:55:59 so its compressing them down to the same range 2011-06-05T14:56:05 maybe 2011-06-05T14:56:08 i guess a lot of good folks don't have bots yet 2011-06-05T14:56:12 that will change soon:) 2011-06-05T14:56:29 bhasker: is a1k0n going to compete? 2011-06-05T14:56:30 well, it also says one should not upload the good bots into the beta for security reasons and such :) 2011-06-05T14:56:45 no idea, he was planning on writing a bot for ants but don't know if he got to it 2011-06-05T14:56:56 fair enough 2011-06-05T14:57:04 anyway, i'm crawling into bed for a few hours, night 2011-06-05T14:57:28 i guess so, but frankly unless someone has a top secret bot, most of the ones right now seem to be very similar 2011-06-05T14:57:44 grab food, run when you can't win, grab more food, fight when you can't do anything else 2011-06-05T14:57:45 i can't imagine people have done much for fighting? 2011-06-05T14:57:53 that's really where i expect the "goodness" to come from with bots 2011-06-05T14:57:58 fighting is mostly restricted to i think don't enter combat if you die 2011-06-05T14:58:07 frankly i think fighting is too simplistic 2011-06-05T14:58:08 right now 2011-06-05T14:58:18 i don't see much strategy happening in fighting 2011-06-05T14:58:33 a bigger fight radius could make fighting better 2011-06-05T14:58:42 i would expect more strategy to come from controlling territory 2011-06-05T14:58:46 rather than the actual fights themselves 2011-06-05T14:58:56 how to position ants to increase ability to gather food 2011-06-05T14:59:02 thats what would decide most games 2011-06-05T14:59:36 actually 2 things a) fast initial growth b) territory control 2011-06-05T14:59:39 that would be key 2011-06-05T14:59:56 fighting is pretty much don't be stupid and walk in to a square where it will put you in range of 2 ants 2011-06-05T15:00:00 compared to yours one 2011-06-05T15:00:38 you could tell where food spawned even if it's outside your fog of war if we implement proper symmetric food 2011-06-05T15:00:59 i think start of food spawn should be symmetric 2011-06-05T15:01:03 after that it should be random 2011-06-05T15:01:13 bhasker: define start 2011-06-05T15:01:13 only the starting food spawn needs to be symmetric 2011-06-05T15:01:19 first turn 2011-06-05T15:01:21 food spawning 2011-06-05T15:01:28 after that its better to be random 2011-06-05T15:01:39 or maybe the first 5-10 turns 2011-06-05T15:01:56 so that no bot has a weird start advantage due to asymmetric food spawns 2011-06-05T15:02:13 but after that random will prevent bots from guessing food locations based on initial conditions 2011-06-05T15:06:48 Why you are against guessing food locations? It coul be another problem to solve by bots. 2011-06-05T15:07:32 i guess i would rather have the bots do some exploration rather than just route to possible food locations straight up 2011-06-05T15:07:37 seems too much like cheating 2011-06-05T15:08:18 and it won't be that hard for most bots to do 2011-06-05T15:08:26 so most smart bots will just go from food location to food location 2011-06-05T15:08:52 it will be rather boring to watch frankly 2011-06-05T15:09:43 i think having a starting similarity surely helps to even the bots ,but beyond that random spawning actually tests the skill of the bot 2011-06-05T15:09:52 to be able to explore and keep control of its existing territory 2011-06-05T15:10:27 otherwise you will have bots that will do stuff like 2011-06-05T15:10:30 rushing 2011-06-05T15:10:38 *** locutus2 has quit IRC (Ping timeout: 252 seconds) 2011-06-05T15:10:54 if i know map is X by Y and there 7 players i can roughly calculate the rstart locations 2011-06-05T15:11:03 a bot could potentially just rush a couple of players with all their ants 2011-06-05T15:11:08 by knowing their start locations 2011-06-05T15:11:20 frankly i believe the bot should not be told how many players are there 2011-06-05T15:11:21 either 2011-06-05T15:11:32 I dont think it is 2011-06-05T15:11:37 otherwise we will see rush strategies like the ones above 2011-06-05T15:11:39 i think it is 2011-06-05T15:12:05 *** berak has joined #aichallenge 2011-06-05T15:12:15 you get enemy ant IDs counting up from 1, 1 being the first one you encounter, not player IDs 2011-06-05T15:12:27 hmm k 2011-06-05T15:12:29 and there is no start param saying you the number of players afaik 2011-06-05T15:12:51 doesn't it send you the scores? 2011-06-05T15:12:59 *** Jaggu has quit IRC (Quit: ChatZilla 0.9.87 [Firefox 4.0.1/20110413222027]) 2011-06-05T15:13:02 after the game is over 2011-06-05T15:13:02 or is it at the end 2011-06-05T15:13:04 ah 2011-06-05T15:13:31 also i think the same map should not be used with same number of players every time 2011-06-05T15:13:41 each map should support 2-max players 2011-06-05T15:13:47