2011-04-18T00:22:56 *** antimatroid has joined #aichallenge 2011-04-18T00:26:20 *** amstan has quit IRC (Read error: Operation timed out) 2011-04-18T00:32:16 *** ltriant has quit IRC (Ping timeout: 246 seconds) 2011-04-18T00:43:29 *** amstan__ is now known as amstan 2011-04-18T01:04:32 *** Mathnerd314 has quit IRC (Ping timeout: 258 seconds) 2011-04-18T01:26:48 *** antimatroid has quit IRC (Ping timeout: 252 seconds) 2011-04-18T01:37:44 *** antimatroid has joined #aichallenge 2011-04-18T01:39:02 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-04-18T01:54:23 *** ltriant has joined #aichallenge 2011-04-18T01:56:26 anyone around? 2011-04-18T01:58:27 mcleopold/antimatroid/amstan? 2011-04-18T01:58:36 hi 2011-04-18T01:58:40 bhasker? 2011-04-18T01:58:49 hey i ran into some weird state 2011-04-18T01:58:54 .a*a. 2011-04-18T01:59:00 basically when my bot got into this state 2011-04-18T01:59:03 no new ant is spawning 2011-04-18T01:59:56 hmm, i'll take a look later 2011-04-18T02:09:36 *** Ice_Harley has joined #aichallenge 2011-04-18T02:17:25 Ice_Harley: I have no way of handling that case fast enough 2011-04-18T02:17:30 and smart enough 2011-04-18T02:17:57 it's a bit of a pitty people wont have enough time to do the really interesting ideas 2011-04-18T02:18:05 but people insist on having larger maps 2011-04-18T02:18:40 which will make language choice even more important 2011-04-18T02:18:54 it's going to be a contest of who can do the best stuff in the time limit 2011-04-18T02:19:05 and macro strategy of course :P 2011-04-18T02:21:02 antimatroid: so you just go to closest food everytime? 2011-04-18T02:21:17 reusing already moved ants yes 2011-04-18T02:21:41 i considered moving to the food square closest to me that minimises the distance to an enemy each iteration, but i don't think that'd work as well 2011-04-18T02:21:43 I didn't find soluthion for this either 2011-04-18T02:23:55 ideally you want to use a game tree for that too 2011-04-18T02:23:59 but i'm saving that for battles 2011-04-18T02:27:59 I thought about trees. What kind of tree you can do? Each node is one move? Or each node is ant from food spawn? 2011-04-18T02:29:34 First variant will have huge number of nodes. Second - how to handle that when one ant takes food others shold go somewhere? 2011-04-18T02:37:19 when i collect food now i already update the information at the food square as to when that will be available for collecting more food 2011-04-18T02:37:33 so yeah, i would be moving it to "next to" the food square in the tree 2011-04-18T02:45:39 but you can't build new position because you don't know where will be other ants on this step 2011-04-18T02:53:49 *** jaycobcoleman has joined #aichallenge 2011-04-18T03:06:06 Ice_Harley: what do you mean? 2011-04-18T03:08:44 Each node in tree is map position, isn't it? 2011-04-18T03:09:41 *** jaycobcoleman has quit IRC (Ping timeout: 252 seconds) 2011-04-18T03:10:23 each node in the tree contains a game state 2011-04-18T03:10:30 each edge is a strategy 2011-04-18T03:11:54 yes-yes. So when you create new node where strategy is ant A1 goes to food F1 you need new game state. In this game state A1 is near F1, and instead F1 you have new ant A2. But where are ants A3, A4 and A5? 2011-04-18T03:12:27 where they were before unmoves 2011-04-18T03:12:48 then for each of those children you move A3, A4 and A5 2011-04-18T03:13:07 or A1 and A2 again for collecting food 2011-04-18T03:14:12 You don't take into account that A1 and A3 are moving simultaneously? 2011-04-18T03:14:23 for blocking paths? 2011-04-18T03:14:24 ARGH, a combination of github's SSL certificate and redirect setup trigger a bug in wget such that you can't securely download tarballs using wget, this is making my life difficult in trying to refactor the worker setup scripts in preparation for adding jailuser chroot setup to them so I can then add that to the sandbox. 2011-04-18T03:14:40 I think it's time to go to bed and work on it tomorrow 2011-04-18T03:14:46 *** jaycobcoleman has joined #aichallenge 2011-04-18T03:15:08 one option is to have a map set > to mark when a square will be "taken" 2011-04-18T03:15:16 and use that information when path finding 2011-04-18T03:15:26 /s/map/set 2011-04-18T03:15:32 or just remove map :P 2011-04-18T03:16:20 int represents the turn it will be occupied on 2011-04-18T03:16:55 but i suspect to be quite limited on time 2011-04-18T03:17:01 so i'm not really sure what will work best yet 2011-04-18T03:17:13 expect* ergh :P 2011-04-18T03:17:43 for blocking pathes I think it's much easier to check moves before submitting to server and if two ants go to one square just recalculate path for one of them, marking that square as water 2011-04-18T03:18:48 information on the location of water squares isn't reset each turn by default, so watch out doing that. 2011-04-18T03:20:36 you can't mark a square as water, you need to retain water information across turns 2011-04-18T03:20:36 there are times where you might have ants cross each others paths that you need to account for 2011-04-18T03:20:36 but you don't want to say the ant can't walk through the path of another ant at all 2011-04-18T03:20:36 that would really block off what you can actually do 2011-04-18T03:20:36 for my squares, i have int ant, and vector newAnts, deadAnts 2011-04-18T03:20:43 jaycobcoleman: it's in temporary game state. I will not change real one 2011-04-18T03:21:08 when i move an ant, i place it onto newAnts (same as when i'm inputting state information), then when i increment states or resolve information etc. I move the ants back from newAnts to ant 2011-04-18T03:22:02 that way in your game tree for example (or heuristics if you don't mind the odd "crash of ants") you can move an ant onto a square containing another ant before it moves 2011-04-18T03:22:13 *** sigh has joined #aichallenge 2011-04-18T03:22:16 and in the case of moving enemies with a game tree, you can move them onto a square containing an enemy etc. 2011-04-18T03:23:18 finished my first day of work at Google :D 2011-04-18T03:23:25 another thing you will want to do at some point is work out how to retain information about food squares that existed when they went out of visibility of you 2011-04-18T03:23:41 sigh: I'm so jealous of you 2011-04-18T03:23:48 the annoying thing is you are only implicitly told if it's no longer there when you get back in view by not being told it's there 2011-04-18T03:23:51 :) 2011-04-18T03:23:53 sigh: :) 2011-04-18T03:23:59 didn't know you were off to google :P 2011-04-18T03:24:36 I don't think I mentioned it in the channel :P 2011-04-18T03:24:37 i was against having implicit information like that, but meh 2011-04-18T03:24:50 my comment was half indicating that :P 2011-04-18T03:25:11 a friend interned there the previous summer and really liked it 2011-04-18T03:25:15 especially the food :P 2011-04-18T03:25:26 haha, yeah... the food is awesome 2011-04-18T03:25:49 what are you doing there? :P 2011-04-18T03:26:13 just normal software engineering role :P 2011-04-18T03:27:02 sigh: What experience do you have that landed you a job with google? I feel like you'd have to have done something pretty damn impressive 2011-04-18T03:27:25 I think it' 2011-04-18T03:27:38 I think it's more that I did quite well in the interviews 2011-04-18T03:27:52 ^^ he knows quite a lot too 2011-04-18T03:28:07 it's all algorithms and system design type questions, so you have to be on top of that 2011-04-18T03:28:24 What did you do before google? 2011-04-18T03:28:26 yeah, i've heard they grill quite well there 2011-04-18T03:28:43 most technical jobs require a pretty good grilling don't they? 2011-04-18T03:28:55 i've been told the RBA will grill you until you don't know what you're talking about 2011-04-18T03:28:59 no matter how long it takes them 2011-04-18T03:29:14 jaycobcoleman: systems developer for a small cloud based mail/web filtering company 2011-04-18T03:30:15 I see. I hope to work at google someday. When I'm a few years older perhaps 2011-04-18T03:30:33 :) 2011-04-18T03:30:47 well if you are talking active interesting in things like the AI challenge, that's a good start :P 2011-04-18T03:31:09 you learn a lot of stuff from participating in things like this 2011-04-18T03:31:28 *active interest 2011-04-18T03:31:42 I've honestly gained 90% of my programming experience in the last month 2011-04-18T03:31:55 are you in uni? 2011-04-18T03:32:00 younger 2011-04-18T03:32:07 I'm 17 2011-04-18T03:32:20 jaycobcoleman: you're doing fine 2011-04-18T03:32:23 ah, I had hardly programmed much at all before uni :) 2011-04-18T03:32:29 *** ltriant has quit IRC (Quit: Computer has gone to sleep) 2011-04-18T03:32:30 i had never written a program before i started uni 2011-04-18T03:32:32 just wrote some stuff for my TI-8s :P 2011-04-18T03:32:36 *TI-83 2011-04-18T03:32:42 and didn't do the top maths in college, or physics and chemistry 2011-04-18T03:32:44 haha I've done lots of that 2011-04-18T03:33:29 when you are in uni try to get as many internships as possible 2011-04-18T03:33:30 jaycobcoleman: do you know of projecteuler.net? 2011-04-18T03:33:39 yes indeed 2011-04-18T03:33:43 i've done no internships :P 2011-04-18T03:34:09 i've had an ra job for over 2 years though, and did a summer research thing 2011-04-18T03:34:10 I did one... wish I had done more 2011-04-18T03:34:16 i'm more on track for academia 2011-04-18T03:34:22 that's sort of the same 2011-04-18T03:34:23 yeah 2011-04-18T03:34:38 i have actually been working this year :P 2011-04-18T03:34:43 :) 2011-04-18T03:35:19 I got offered an internship with linden air cargo today, which I guess relates somehow to programming? I don't know 2011-04-18T03:35:28 did you mention the ai challenge when applying/interviewing? 2011-04-18T03:35:33 i listed it for gsoc :P 2011-04-18T03:35:48 *** mceier has joined #aichallenge 2011-04-18T03:35:49 I didn't actually apply 2011-04-18T03:35:56 so no 2011-04-18T03:36:02 i was talking to sigh :P 2011-04-18T03:36:15 There we go. Makes more sense 2011-04-18T03:36:17 antimatroid: I didn't list it on my resume... I mentioned it to some of the interviewers 2011-04-18T03:36:52 my projecteuler record is more noteworthy than my aichallenge record though :P 2011-04-18T03:37:00 did you mention that? 2011-04-18T03:37:02 i don't mention that 2011-04-18T03:37:08 although i've only done 50 something problems 2011-04-18T03:37:10 I mentioned that 2011-04-18T03:37:12 and haven't touched it in ages 2011-04-18T03:37:22 well, no one asked me how many problems I had done 2011-04-18T03:37:30 animatroid: Is project euler really worth looking into? Cause I just came across it somewhat recently and haven't really gone back to it 2011-04-18T03:38:03 jaycobcoleman: it's useful for learning a language, and picking up some skill with problems like that 2011-04-18T03:38:04 and a bit of fun 2011-04-18T03:38:11 jaycobcoleman: I would highly recommend it, especially if you want to explore the maths side more 2011-04-18T03:38:30 I do like me the maths side 2011-04-18T03:38:42 jaycobcoleman: if you are serious about cs and not just programming, i can't recommend doing a dual maths major enough :) 2011-04-18T03:39:14 and i'm not very knowledgeable about cs, but i sort of know about the different things :P 2011-04-18T03:39:55 in cs if you know the basic concepts it's much easier to pick up anything else 2011-04-18T03:40:15 i think that's the case for everything :P 2011-04-18T03:40:33 Well then, to problem number one it is 2011-04-18T03:40:51 antimatroid: yes and no 2011-04-18T03:40:51 what language do you use? 2011-04-18T03:41:12 antimatroid: for example in math, ever new subject redefines what is "basic" 2011-04-18T03:41:48 queue hitler topolgy video :P 2011-04-18T03:41:53 maybe, it's still good to know the basics 2011-04-18T03:42:02 i love that video 2011-04-18T03:42:05 what do you count as the basics for maths? 2011-04-18T03:42:27 depends on what areas of maths, is there another word that encompasses more? 2011-04-18T03:42:32 everything :P 2011-04-18T03:42:36 but you know what i mean 2011-04-18T03:42:53 that was my point, the basics covers too much :P 2011-04-18T03:42:58 for pure maths say, knowing about sets, relations, etc. is vital 2011-04-18T03:43:09 then you just keep building 2011-04-18T03:44:24 animatroid: Where you asking what language I use? 2011-04-18T03:44:56 he was 2011-04-18T03:45:12 mostly python 2011-04-18T03:45:16 and C++ 2011-04-18T03:45:21 I know you like that one 2011-04-18T03:47:42 anything but java and i wont hate on you too much :P 2011-04-18T03:48:15 imo c++ is better for maths, because if you define matrices you can overload operators on them 2011-04-18T03:48:17 I don't understand java worth crap 2011-04-18T03:49:07 Matrix A = 2*B*(C^4) is a lot better than 2011-04-18T03:49:07 Matrix A = smult(2, mult(B, pow(C, 4))) 2011-04-18T03:49:20 antimatroid: you can do that in python too 2011-04-18T03:49:24 ah, nevermind then 2011-04-18T03:49:46 infact, numpy already does that for you :P 2011-04-18T03:50:01 does it do tensors? 2011-04-18T03:50:06 I like numpy :) 2011-04-18T03:50:45 antimatroid: dunno, haven't worked with tensors 2011-04-18T03:50:55 i am tempted to do a second normal form game class with utilities represented as map 2011-04-18T03:51:04 it's slower, but way easier 2011-04-18T03:51:38 I'm using numpy to keep track of the soonest that an enemy ant can be on each square given the information I have 2011-04-18T03:52:02 jaycobcoleman: how are you doing that? 2011-04-18T03:53:46 everytime I get vision information from my ants I'll assign the value of the vision radius to the square I'm on and then a value one less for each square away from that point, then at the beginning of each turn I'll subtract one from the entire array 2011-04-18T03:54:36 a good way would be as follows: create a "time accessible map" set at -1, add all enemy locations to a queue noting the current distance for each location (currently 0) and set those locations as 0 in the accessible map, then while the queue isn't empty, for any adjacent square with accessible at -1, update that to the current squares value + 1, and add that information to the end of the queue 2011-04-18T03:55:11 then your accessible map will tell you the shortest turns in which an enemy can reach any square on the map 2011-04-18T03:55:34 you can do that from food squares to get closest ant to food, from all ants to partition the map into who can reach what first, etc. etc. 2011-04-18T03:56:05 Yes I'm planning on doing something similar 2011-04-18T03:56:12 I haven't worked out all the logistics 2011-04-18T03:57:51 But I plan to come up with several possible moves for each of my ants and evaluating the likelyhood of their success and then any conflict between two ants regarding movement onto a particular aquare will be resolved by the likelyhood of success 2011-04-18T03:57:58 *square 2011-04-18T03:58:21 yep, i'm planning to do that just for ants that may land in a battle 2011-04-18T03:58:24 and use a game tree for it 2011-04-18T03:58:40 i expect time to be tight 2011-04-18T03:59:01 I know an unfortunately small amount about game theory 2011-04-18T03:59:59 the basics are quite simple 2011-04-18T04:00:07 there's a bit of jargon one needs to get used to though 2011-04-18T04:00:11 also, I feel like time shouldn't be a concern in this contest, to a reasonable extent anyway. 2011-04-18T04:00:20 it will be 2011-04-18T04:00:27 and unfortunately it'll be necessary 2011-04-18T04:00:41 with a few thousand bots, we need quite a few games for decent rankings, and the contest can't run forever 2011-04-18T04:01:00 WE NEED MORE SERVERS!!! 2011-04-18T04:01:14 that requires moar money 2011-04-18T04:01:49 psh, It's sponsored by google 2011-04-18T04:02:06 they've got loads of money 2011-04-18T04:05:00 they have given a bit of money previously, and hopefully some this time 2011-04-18T04:05:15 but the names mostly there because it looks good for both parties 2011-04-18T04:05:34 Yeah I kinda figured that out by this point 2011-04-18T04:06:04 it would be silly to remove it unless google actually wanted us to / stopped giving permission 2011-04-18T04:07:53 the contest is still quite new, but it's grown well, there was ~800 bots for tron, then ~4700 for planet wars 2011-04-18T04:08:08 the contests before that were just for uwaterloo students 2011-04-18T04:08:21 rock, paper, scissors? 2011-04-18T04:08:21 amstan won the repeated rock paper scissors one 2011-04-18T04:08:31 repeated 2011-04-18T04:08:34 have you been here the while time? 2011-04-18T04:08:38 the bots would play a bunch of games against each other 2011-04-18T04:08:44 *whole 2011-04-18T04:08:46 i came in for tron 2011-04-18T04:08:56 waterloo is in canada, i'm in tasmania, australia :p 2011-04-18T04:09:12 *** amstan has quit IRC (Read error: Operation timed out) 2011-04-18T04:09:15 yes, of course 2011-04-18T04:11:41 *** FireFly has quit IRC (Quit: swatted to death) 2011-04-18T04:15:12 *** Naktibalda has joined #aichallenge 2011-04-18T04:36:19 animatroid: Thanks a lot, now instead of making my bot I'm busy finding palindromes :P 2011-04-18T04:40:04 *** antimatroid has quit IRC (Ping timeout: 252 seconds) 2011-04-18T04:42:50 *** antimatroid has joined #aichallenge 2011-04-18T04:45:15 *** chris__0076 has joined #aichallenge 2011-04-18T04:47:01 *** Chris_0076 has quit IRC (Ping timeout: 240 seconds) 2011-04-18T04:47:18 *** antimatroid has quit IRC (Ping timeout: 252 seconds) 2011-04-18T04:48:24 *** antimatroid has joined #aichallenge 2011-04-18T05:13:36 *** woudshoo has joined #aichallenge 2011-04-18T05:33:26 *** jaycobcoleman has quit IRC (Ping timeout: 252 seconds) 2011-04-18T05:47:35 *** berak has joined #aichallenge 2011-04-18T05:53:10 *** needsch has joined #aichallenge 2011-04-18T05:59:20 *** smiley1983 has joined #aichallenge 2011-04-18T06:13:15 smiley1983: did anyone ever respond about putting the OCaml bot up? 2011-04-18T06:13:34 if the actual functionality is solid i'd say put it up 2011-04-18T06:18:19 sigh: oh yeah 2011-04-18T06:18:30 bhasker found a possible bug with ant spawning 2011-04-18T06:18:40 what was that? 2011-04-18T06:18:41 something about ...a*a... not spawning an ant 2011-04-18T06:18:56 i haven't got around to testing that myself yet 2011-04-18T06:19:28 if you can make a repeatable test case, I'll have a look at it 2011-04-18T06:19:46 i'll have a look in a bit 2011-04-18T06:50:50 *** delt0r_ has joined #aichallenge 2011-04-18T06:52:12 *** delt0r___ has quit IRC (Ping timeout: 246 seconds) 2011-04-18T06:52:26 *** sigh has quit IRC (Ping timeout: 240 seconds) 2011-04-18T07:01:32 antimatroid: no, but unless someone's in a hurry to get their hands on the code, I'll probably finish it first. I've got the git repository cloned, but I haven't done the research into how to make a pull request yet. 2011-04-18T07:01:50 yep, there's definitely no rush 2011-04-18T07:02:01 ps bring it on for first in aus again :P 2011-04-18T07:02:23 The basic functionality is (I believe) solid, though. 2011-04-18T07:02:43 Congratulations on being first last time, and by such a large margin. :) 2011-04-18T07:02:43 i take it you're planning to participate? 2011-04-18T07:02:49 Yes. 2011-04-18T07:02:53 you were worrying me towards the end 2011-04-18T07:02:58 it wasn't clear to me who was best on tcp 2011-04-18T07:03:02 (i didn't use antimatroid on tcp) 2011-04-18T07:03:23 but thanks :) 2011-04-18T07:03:33 Ha. I found it difficult to judge much from the TCP games because the pool of regular opponents was so small 2011-04-18T07:03:49 still better than testing on the main site 2011-04-18T07:03:57 i didn't even bother with a bot on there for the last month 2011-04-18T07:03:59 Oh yes, it was a great blessing 2011-04-18T07:04:20 That'll be why I didn't notice you among the top ranks until the very end. :) 2011-04-18T07:04:37 yep, i'm tempted to do the same again 2011-04-18T07:04:42 it's nice jumping in at the end 2011-04-18T07:05:15 Can anyone tell me, for the starter pack: should the distance function return the number of steps an ant has to take, or the LOS distance? 2011-04-18T07:05:18 the plan is to make tcp more official this time, so hopefully we can get more people testing on there 2011-04-18T07:05:28 LOS? 2011-04-18T07:05:30 That would be great. 2011-04-18T07:05:33 Line of Sight 2011-04-18T07:05:36 distance is the euclidean distance with edges wrapped 2011-04-18T07:05:41 check the specs page 2011-04-18T07:06:11 https://github.com/aichallenge/aichallenge/wiki/Ants-Game-Specification 2011-04-18T07:06:47 all distances are using that metric 2011-04-18T07:07:13 which is unfortunately tricky to work with when you do vision calculations, be really careful there 2011-04-18T07:07:39 ight. I think the python starter bot might be returning number of steps 2011-04-18T07:07:47 if you do a simple bfs from all ants simultaneously it doesn't always work properly, you miss some 2011-04-18T07:07:54 have you looked at my c++ bot? 2011-04-18T07:08:02 that's pretty solid, and reasonably well laid out 2011-04-18T07:08:42 I haven't - is it the c++ starter bot or available elsewhere? 2011-04-18T07:09:27 the c++ starter bot 2011-04-18T07:09:27 also - bfs? 2011-04-18T07:09:33 breadth first search 2011-04-18T07:09:45 basically, add all ants locations to a queue, then keep "painting outwards" 2011-04-18T07:10:29 it's nice for doing stuff like partitioning the squares into which player can reach them first, or finding which ant is closest to a food square etc. 2011-04-18T07:11:54 Yes. I'm looking at the c++ code now - I had used the Python one for reference because Python is readable, but I've noticed some incompletenesses. I'll use yours for reference from now, thankyou for pointing me to it 2011-04-18T07:14:32 antimatroid did you test my case? 2011-04-18T07:14:57 bhasker: not yet sorry, got a few other things to do 2011-04-18T07:15:09 kk 2011-04-18T07:16:34 i implemented a basic BFS search to find closest ant 2011-04-18T07:16:38 and extract the path 2011-04-18T07:16:47 to issue orders 2011-04-18T07:16:49 but... 2011-04-18T07:17:03 that bug is stopping my bot from going very far:p 2011-04-18T07:18:39 strange, i ran games which flooded the map and never noticed anything like that 2011-04-18T07:20:12 i can send you a text copy of the map state if you want 2011-04-18T07:20:16 or a replay 2011-04-18T07:22:32 do it to sigh 2011-04-18T07:22:38 but he's not here 2011-04-18T07:22:49 was just reading the logs he is close to where i am 2011-04-18T07:22:52 now that he works at google 2011-04-18T07:23:03 except he's in sydney :P 2011-04-18T07:23:07 oooh 2011-04-18T07:23:08 doh 2011-04-18T07:23:18 i thought he joined the US office 2011-04-18T07:23:24 nup 2011-04-18T07:23:56 smiley1983: what state are you in? 2011-04-18T07:27:42 *** sigh has joined #aichallenge 2011-04-18T07:28:16 there you go 2011-04-18T07:29:50 I'm in Sydney too 2011-04-18T07:30:27 meh, tassie is better :P 2011-04-18T07:31:24 Yes, I would consider living there if I wasn't set up here 2011-04-18T07:31:47 i'll probably have to move at the end of this year 2011-04-18T07:31:53 but i always see myself ending up back here 2011-04-18T07:32:23 Do you like bushwalking? 2011-04-18T07:32:39 yeah and camping 2011-04-18T07:32:46 particularly camping and drinking 2011-04-18T07:33:05 Definitely a great place yo live then 2011-04-18T07:33:40 my only complaint is nothing in the way of snowboarding/skiing 2011-04-18T07:33:53 i've always wanted to, but only ever snowboared for one day 2011-04-18T07:34:52 Ah. I've been assured it's fun, but it just doesn't appeal to me for some reason. 2011-04-18T07:35:34 i used to skate a fair bit, so it's always seemed like it'd be even better 2011-04-18T07:35:50 *** berak has quit IRC (Quit: ChatZilla 0.9.84 [SeaMonkey 2.0a3/20090223135443]) 2011-04-18T07:36:40 Requires more equipment, and big hills covered in snow, though 2011-04-18T07:37:07 yeah that's fine 2011-04-18T07:37:19 i'm tempted to move somewhere that's it's easily accessible 2011-04-18T07:37:47 Yes, it makes a great deal of sense in snowy places :) 2011-04-18T07:38:38 Do you have any idea of the expected duration of this contest? 2011-04-18T07:38:54 the number i proposed was 2 months 2011-04-18T07:39:03 because tron was too short and planet wars too long 2011-04-18T07:39:14 but there's going to be a beta period for like 2 weeks as well, for ironing stuff out 2011-04-18T07:39:36 although i'm tempted to change my opinion back to 3 months 2011-04-18T07:39:44 so people who come in have plenty of time 2011-04-18T07:41:21 Right. I was thinking it might be good to update the front page so casual visitors know something is happening without checking the forums 2011-04-18T07:42:04 I checked in sometime in March but didn't look at the forums, and thought nothing was happening yet. 2011-04-18T07:42:16 hmmm maybe 2011-04-18T07:42:21 I only checked the forums a few days ago and found that the next game had been determined 2011-04-18T07:42:22 i'll mention that to amstan 2011-04-18T07:42:35 yeah, we've been working for ages 2011-04-18T07:42:50 have you taken a look at the symmetric maps? 2011-04-18T07:43:00 you can turn symmetric food spawning on too 2011-04-18T07:43:32 Yes, the test map I've been using is symmetric 2011-04-18T07:43:43 I haven't fiddled with any options, though 2011-04-18T07:44:32 --food symmetric 2011-04-18T07:44:41 add that to your play_one_game script to do symmetric food 2011-04-18T07:44:52 feel free to make another map generator, we want lots of variety 2011-04-18T07:45:28 Right. I'll be looking at other things to develop once I'm happy (enough) with my starter pack. :) 2011-04-18T07:47:15 are you watching games with the viewer or visualiser? 2011-04-18T07:47:21 the visualiser is far superior 2011-04-18T07:47:48 Visualiser - I hadn't noticed the viewer 2011-04-18T07:48:03 cool, ignore what i said 2011-04-18T07:48:56 So far there isn't anything to watch, though, except to confirm that it's making the correct starter bot moves. 2011-04-18T07:49:25 yeah okay 2011-04-18T07:49:34 i've played around with food collection, which is fun to watch 2011-04-18T07:49:54 antimatroid another bug in the c++ starter bot 2011-04-18T07:50:00 it doesn't clear out old food in the grid 2011-04-18T07:50:00 :( 2011-04-18T07:50:07 ah my bad 2011-04-18T07:50:14 so if the food is now an ant 2011-04-18T07:50:20 it still shows up as a food etc 2011-04-18T07:50:32 so probably in the loop which reads in the state 2011-04-18T07:50:36 you need to validate all the food 2011-04-18T07:50:37 yes it does 2011-04-18T07:50:45 well, it should :\ 2011-04-18T07:50:55 clear() in square.h has isFood = 0 2011-04-18T07:51:07 it won't right since food information is sent only once when its seen first time? 2011-04-18T07:51:11 and you have this in your loop 2011-04-18T07:51:22 for(int row=0; row for(int col=0; col if(!grid[row][col].isFood) 2011-04-18T07:51:23 grid[row][col].reset(); 2011-04-18T07:51:29 so it doesn't reset Food Squares 2011-04-18T07:51:37 oh, that should be isWater 2011-04-18T07:51:38 cheers 2011-04-18T07:51:53 will food information be sent every turn? 2011-04-18T07:51:59 all visible food? 2011-04-18T07:52:02 if it's visible 2011-04-18T07:52:04 ah 2011-04-18T07:52:28 but if you go out of range of food while it's there, you're only implicitly told it's not there anymore when you come back in range by not being told it's there 2011-04-18T07:52:31 water squares arent sent right 2011-04-18T07:52:38 i really don't like that, but i didn't win that battle 2011-04-18T07:52:49 water square information is sent once per bot 2011-04-18T07:52:56 doh fixing that makes my bot behave better 2011-04-18T07:52:57 lol 2011-04-18T07:53:02 that was throwing my bot off 2011-04-18T07:53:06 my ants will step on each other 2011-04-18T07:53:11 and die 2011-04-18T07:53:41 That reminds me - is it okay for a starter pack to track the turn on which each tile was last seen (for the Visible function) or should it clear this info between turns? It's obviously useful to keep it, but not specified in the starter pack guide. 2011-04-18T07:55:54 aichallenge: Nick Ham epsilon * rad32822 / ants/dist/starter_bots/cpp/State.cc : fixed a food information bug - http://bit.ly/fBTQkW 2011-04-18T07:56:30 there we go :) 2011-04-18T07:56:56 smiley1983: i think i am? 2011-04-18T07:57:07 :) Right. 2011-04-18T07:57:15 hah BFS doesn't scale so well once the search space increases:-\ 2011-04-18T07:57:22 Found food: 105,62 2011-04-18T07:57:22 Issuing order : 102,43,E 2011-04-18T07:57:22 time taken: 806.759ms 2011-04-18T07:57:24 :-\ 2011-04-18T07:57:42