2011-04-03T00:00:18 people don't need to worry about programming for such a special case and not ending the game if it would result in a lower score 2011-04-03T00:00:27 or a higher score for the enemy.. 2011-04-03T00:00:29 potentially.. 2011-04-03T00:00:50 the idea is you would still want to go in for the kill 2011-04-03T00:01:20 i know it's a very unlikely edge case, but it's fairly easy to say the food bonus goes to the last participating player if there is one and possibly works better? :P 2011-04-03T00:03:11 sure 2011-04-03T00:03:41 hmm.. at the moment I think I've just got turnsremaining*spawnrate so it won't do anything anyway 2011-04-03T00:04:37 ... you should be adding the number of food squares currently on the grid too :) 2011-04-03T00:04:40 that will fix it 2011-04-03T00:05:11 it would likely be a pretty small bonus unless both armies are small 2011-04-03T00:05:42 *** stocha has joined #aichallenge 2011-04-03T00:05:50 both armies? 2011-04-03T00:05:52 what memory can we expect to have ? 2011-04-03T00:06:09 well, the dying army has to be small, the winning army might be small too 2011-04-03T00:06:26 stocha: wont know until we work out how everything is to be run 2011-04-03T00:06:32 i'm not banking on much if we do multiplayer 2011-04-03T00:07:17 will it be more like 1kb or more like 1gb ? 2011-04-03T00:07:36 or somewhere in between ? 2011-04-03T00:08:28 i realize i'm trying to optimize memory. And that could be so wrong :p 2011-04-03T00:08:32 i would say 200mb-1gb would be my guess 2011-04-03T00:08:41 okay 2011-04-03T00:08:44 maybe less 2011-04-03T00:08:50 anyone think i can bet on 100 mo ? 2011-04-03T00:08:51 1gb per bot seems excessive 2011-04-03T00:08:55 are you doing a tree already? 2011-04-03T00:09:07 no ... but still i'm worrying :p 2011-04-03T00:09:21 i'm worried about my tree, i may have to clear state information as i generate 2011-04-03T00:09:26 i'd like to maintain some full int structure for each square * number of entity 2011-04-03T00:09:39 (ant food, maybe even water i don't know) 2011-04-03T00:09:45 it is starting to add up :p 2011-04-03T00:10:12 i have a fair bit of information for each square, and i have a grid of those, which each node in my tree will have 2011-04-03T00:10:22 but i was like maybe we will have only 1mb. 2011-04-03T00:10:28 i expect 100mb to be more than enough in the worst case where i clear data as i generate 2011-04-03T00:10:52 if 100 mb is okay i'll be fine up until i do trees i guess. 2011-04-03T00:10:54 you shouldn't be getting less than 100mb i don't think 2011-04-03T00:10:55 i hope. 2011-04-03T00:11:02 what will the max map size be ? 2011-04-03T00:11:03 stocha: some of the language run times won't fit in 1 mb :P 2011-04-03T00:11:15 stocha: will depend on what we can visualise 2011-04-03T00:11:22 people are keen for as large as possible i think 2011-04-03T00:11:24 yeah but that doesn't say much about the memory of the structures :p 2011-04-03T00:11:44 maybe i should drop the grid representation 2011-04-03T00:11:52 yeah but with my idea of using int map based on real map size (for simplicity really) i have to know a size limit :p 2011-04-03T00:12:13 there will be a limit when the contest starts 2011-04-03T00:12:20 or soon after 2011-04-03T00:12:23 if it gets much larger than 128 i'll have to start worrying 2011-04-03T00:12:26 until then assume anything 2011-04-03T00:12:38 if it is larger then 512, then i have to really take it into account 2011-04-03T00:12:42 i doubt we'd get over 200x200 and i personally don't think that's feasible 2011-04-03T00:13:12 we need to also take into consideration the length of games, and allowing games to run long enough to actually play out 2011-04-03T00:13:13 okay so 256x256 will be the definite limit then ? 2011-04-03T00:13:25 200x200 is hard to show and would require fairly lengthy games 2011-04-03T00:13:33 good 2011-04-03T00:13:33 stocha: i can give you no definite limit 2011-04-03T00:13:42 have you any authority over the case ? 2011-04-03T00:13:42 i am strictly giving you my current opinion on the topic 2011-04-03T00:13:46 no i don't 2011-04-03T00:14:10 had you some discussion with any authorities about it ? :p 2011-04-03T00:14:28 i'm fairly confident jeff will agree with what i just said 2011-04-03T00:14:36 Still i suppose most will agree that 200+ isn't that good :p 2011-04-03T00:15:03 okay, so i have 100mb and 256x256 max size :p 2011-04-03T00:15:27 i think you would be fairly safe taking those as worst case scenarios for the moment 2011-04-03T00:15:35 you and i both will need to reevaluate a lot of things otherwise 2011-04-03T00:15:51 mmm i need the length of game and max number of food per game :p 2011-04-03T00:16:03 sigh: what's the current spawn 2011-04-03T00:16:10 we can't lock a length in now 2011-04-03T00:16:15 current spawn? 2011-04-03T00:16:19 rate 2011-04-03T00:16:21 give some max value again 2011-04-03T00:16:30 there should be pretty safe guess. 2011-04-03T00:16:33 1 food per turn per player 2011-04-03T00:16:40 no starting food ? 2011-04-03T00:16:40 length will depend on what allows for a game to play out with the dimensions that are decided upon 2011-04-03T00:16:52 there is starting food 2011-04-03T00:17:00 ah yeah, i was going to say otherwise :P 2011-04-03T00:17:05 what will be the limit for starting food ? 2011-04-03T00:17:07 but I would not count on any of that staying the same 2011-04-03T00:17:16 stocha: not part of the specs 2011-04-03T00:17:19 sigh: there should always be starting food 2011-04-03T00:17:36 antimatroid: sure, but the amount of starting food is not part of the specs 2011-04-03T00:17:38 i still need a upper limit, for anticipating worst case memory usage. 2011-04-03T00:17:49 stocha: upper limit is all land 2011-04-03T00:17:58 65535 ! 2011-04-03T00:18:00 all food* 2011-04-03T00:18:05 that's a lot of ant to care for ! 2011-04-03T00:18:14 *all land is food :P 2011-04-03T00:18:20 we wont flood the maps, but you are asking for an upper bound 2011-04-03T00:18:21 but then the food per turn would drop i guess :p 2011-04-03T00:18:22 stocha: you asked for an upper limit 2011-04-03T00:18:22 that is one 2011-04-03T00:18:30 point is that it is not defined 2011-04-03T00:18:34 that's hell of a upper one :p 2011-04-03T00:18:49 and it won't ever be part of the specs 2011-04-03T00:19:00 so i have to account for 65000 ant 2011-04-03T00:19:14 when i would like to go for like max 2048 :p 2011-04-03T00:19:26 you can unofficially bank on the maps not being completely flooded as that would be stupid 2011-04-03T00:19:32 stocha: no, you don't... up to you to decide what you account for 2011-04-03T00:19:42 we're just not willing to give a realistic upper bound, that's the point 2011-04-03T00:19:43 :'( 2011-04-03T00:19:58 but you can't rely on there being a fixed answer 2011-04-03T00:20:27 how are people to make bots if they have no informations on anything :p 2011-04-03T00:20:40 stocha: use intelligence :P 2011-04-03T00:20:43 the most lucky one will win. because his bot won't crash. 2011-04-03T00:20:44 sigh: might be worth playing around with n in {1,2,3} food squares per player each turn 2011-04-03T00:20:47 that's what it is called AI 2011-04-03T00:21:06 it's not ai, it's memory management guessing. 2011-04-03T00:21:25 it wont be that absurd so as to break lots of bots 2011-04-03T00:21:30 stocha: well then wait until the contest has started and the engine is stable 2011-04-03T00:21:36 just don't restrict yourself to a really small number of ants per team 2011-04-03T00:21:43 ah you will provide information then ? :) 2011-04-03T00:21:56 the engine code will be open source 2011-04-03T00:21:57 i was started to worry that no upper bound was the spec. 2011-04-03T00:22:02 stocha: the engine source code is always available 2011-04-03T00:22:11 it won't change during the ongoing contest ? 2011-04-03T00:22:26 it won't change in the finals 2011-04-03T00:22:28 we're hoping to lock things down like that earlier rather than later 2011-04-03T00:22:36 but no absolute guarantee 2011-04-03T00:22:48 the only reason it will change is for bugs or to balance the game 2011-04-03T00:22:53 before the finals 2011-04-03T00:22:59 i still have trouble understanding how decision are made :p 2011-04-03T00:23:22 decisions are made to try and make the most interesting game both visually and strategically and to allow people to get involved easily 2011-04-03T00:23:24 stocha: ok, ask again after the submissions are open 2011-04-03T00:23:28 at least those are my main prioirites 2011-04-03T00:23:29 that's the answer 2011-04-03T00:23:33 it's pointless asking now 2011-04-03T00:24:00 we don't even have an official battle resolution and you want the definite upper bound on maximum starting food squares? :P 2011-04-03T00:24:04 yeah i accept the answer of 1 food per ground square. 2011-04-03T00:24:34 but now i ask this new question about my wondering how decision are settled. Not why (to make the game more interesting) but How. 2011-04-03T00:24:58 jeff started this, he has the ruling opinion 2011-04-03T00:25:22 So decision are made, whenever jeff state that something should be ? 2011-04-03T00:25:22 he is at least willing to listen to peoples opinions and while not necessarily always agreeing with me has a somewhat valid reason for his opinion 2011-04-03T00:25:26 so it could be worse 2011-04-03T00:25:52 hey i just want to know the rules. I guess now i have to try to know jeff :p 2011-04-03T00:26:04 Will he be making a bot also ? 2011-04-03T00:26:21 it has been said over and over again, the rules are not decided and anyone playing around atm should not expect them to be 2011-04-03T00:26:29 what langage is he comfortagle with ? 2011-04-03T00:26:34 he is not allowed to compete, he has server access 2011-04-03T00:26:52 stocha: anyone who has access to the contest servers can't compete 2011-04-03T00:26:55 so you guys antimatroid, mcleopold ans such won't have ? 2011-04-03T00:27:05 no 2011-04-03T00:27:26 but will he still make some bots ? 2011-04-03T00:27:32 even if he doesn't compete ? 2011-04-03T00:27:34 no, he is quite busy 2011-04-03T00:27:47 it wouldn't be as fun when you can't submit it to the final contest 2011-04-03T00:28:01 so he will rely on his lieutenants big time :p 2011-04-03T00:28:01 you will have a few thousand other people making bots, i wouldn't sweat it 2011-04-03T00:28:27 do you guys have been knowing him for long ? 2011-04-03T00:29:02 that's assuming the challenge will encounter success :p 2011-04-03T00:29:16 wich probably every people here at the moment hope for. 2011-04-03T00:30:08 mm, okay that's only about 4 gb of memory for the basic structures. 2011-04-03T00:30:34 so i'm stuck with a dice roll i guess :p 2011-04-03T00:30:56 let's try with ant <= 2048 then 2011-04-03T00:31:59 that would give about 140 mb. 2011-04-03T00:34:31 antimatroid: regarding food bonus, why won't it result in a stalemate when there are more than 2 players? 2011-04-03T00:34:42 sigh: i thought of that 2011-04-03T00:34:44 what then? 2011-04-03T00:35:19 hmm? 2011-04-03T00:35:20 i think really in both situations you both want to be in a fight if possible anyway, to score some points 2011-04-03T00:35:31 maybe without the food bonus in both instances? 2011-04-03T00:35:36 why add a food bonus then ? 2011-04-03T00:35:42 both situations? 2011-04-03T00:35:58 ie. maybe the food bonus shouldn't be given out at the end ever, 2 players or n players 2011-04-03T00:36:52 that way we can just have: if the game ends before the maximum number of turns, the winner gets a food bonus which is equal to the remaining number of turns * spawn rate 2011-04-03T00:37:28 ok, why won't that result in a stalemate? 2011-04-03T00:37:47 you give +1 if it does ? 2011-04-03T00:37:56 if you can get an ant involved in killing another your points increase 2011-04-03T00:37:57 to the remaining guy 2011-04-03T00:38:22 you may not want to fight if you think you are really close on points and not sure who would win 2011-04-03T00:38:27 so the game is about having ant involved in killing one. 2011-04-03T00:38:29 but i think that's the same with the food bonus? 2011-04-03T00:38:48 really you're going to want to try and score any additonal points 2011-04-03T00:39:01 why ? if you are already winning 2011-04-03T00:39:06 so you are arguing that the food bonus is too small to matter? 2011-04-03T00:39:08 you don't know that 2011-04-03T00:39:10 you only have to wait for the game to end. 2011-04-03T00:39:14 ah true. 2011-04-03T00:39:18 but you can guess 2011-04-03T00:39:20 sigh: yeah pretty much 2011-04-03T00:39:28 then why have it? :P 2011-04-03T00:39:30 if you can guess how many ant there is in in the game. 2011-04-03T00:39:36 sigh: and that the food bonus is meant to encourage you to end the game earlier 2011-04-03T00:39:39 the game is over regardless 2011-04-03T00:39:57 if it is too small to matter, why would it encourage the game to end? 2011-04-03T00:40:04 defining the food bonus as above, winning there gives you a food bonus, but it's 0 2011-04-03T00:40:04 so the question is : without food bonus will the game take longer. 2011-04-03T00:40:22 sigh: we're talking about the turn right as we hit max turns 2011-04-03T00:40:28 the game is called off anyway 2011-04-03T00:40:38 stop making that stupid distinction 2011-04-03T00:40:47 ? 2011-04-03T00:40:47 why not state that if the game end before turn limit the remaining player always win ? 2011-04-03T00:40:54 there is no difference between a game that ends early and otherwise 2011-04-03T00:41:05 nor should there be 2011-04-03T00:41:11 it just complicates the discussion 2011-04-03T00:41:26 my point above was there doesn't need to be 2011-04-03T00:41:52 all you need to say is: if the game ends and there is one participating player remaining, they get a food bonus which is the number of turns remaining * spawn rate 2011-04-03T00:42:17 if num turns = 0, the food bonus is nothing 2011-04-03T00:42:35 yes, there has never been an argument about that? 2011-04-03T00:42:35 *** amstan_ has joined #aichallenge 2011-04-03T00:42:35 *** ChanServ sets mode: +o amstan_ 2011-04-03T00:43:07 I don't see why you ever made that distinction 2011-04-03T00:43:19 i started this conversation talking about the food bonus as that specific point :\ 2011-04-03T00:43:45 what specific point? 2011-04-03T00:43:52 end of turn limit 2011-04-03T00:45:08 why does that case matter, or why did that case ever matter as a distinct case? 2011-04-03T00:45:10 won't scoring dead ants increase the incentive for defending positions ? 2011-04-03T00:45:45 sigh: my original point was wanting to avoid a stalemate there with the current metho 2011-04-03T00:45:54 because attacking my earn a lower ratio 2011-04-03T00:45:55 then i thought we came around to the conclusion that i was wrong 2011-04-03T00:46:00 antimatroid: I thought food bonus was always: "turns remaining * spawn rate" or "turns remaining * spawn rate + current food" 2011-04-03T00:46:43 sigh: i wasn't aware that the end case had been considered, i'm sure mcleopold and i have talked about it before 2011-04-03T00:46:55 probably more me querying him on it like this though 2011-04-03T00:47:00 the end case should never have to be considered! 2011-04-03T00:47:15 it should follow as a trivial extension of the general case 2011-04-03T00:47:30 what make you think that the food bonus will change how people tune their bots. 2011-04-03T00:47:38 I'm confused as to why there was ever a distinction! 2011-04-03T00:48:12 if couting killed ant is fair. Then what use for a food bonus ? 2011-04-03T00:50:27 *** Mathnerd314 has quit IRC (Read error: Connection reset by peer) 2011-04-03T00:50:44 antimatroid: anyway, do you agree that there is no need to treat the turnlimit as a special case? 2011-04-03T00:52:46 *** stocha has quit IRC (Ping timeout: 252 seconds) 2011-04-03T00:54:25 meh, I have a feeling we've been talking past each other... I guess I don't understand the terminology you guys have been using 2011-04-03T01:33:51 sigh: i agree with that 2011-04-03T01:34:02 :) 2011-04-03T01:34:44 antimatroid: ok, to clarify the next point, the current consideration is between: "turns remaining * spawn rate" and "turns remaining * spawn rate + current food"? 2011-04-03T01:35:13 i don't mind :P 2011-04-03T01:35:31 what's the current consensus? 2011-04-03T01:35:50 it was with + current food 2011-04-03T01:35:56 but i don't think it's really been discussed 2011-04-03T01:36:13 i see 2011-04-03T01:36:48 and what change were you proposing to avoid stalemate? 2011-04-03T01:37:47 i wasn't sure whether the food bonus was defined at the end of the game like that 2011-04-03T01:38:27 it is the obvious way to do it, you're right, but there's a lot of things that i think are obvious that apparently aren't 2011-04-03T01:38:33 like using row/col for indexing :P 2011-04-03T01:38:36 :P 2011-04-03T01:39:34 either way, I'm unsure how clarifying the case where there is exactly 1 player at the turn limit helps... that is such a rare edge case? 2011-04-03T01:40:35 I guess my question is, were you brining it up for completeness sake, or because it would change the outcome fundementally? 2011-04-03T01:48:06 completeness 2011-04-03T01:48:14 i always bring up stupid edge cases like that :P 2011-04-03T01:48:27 i get ripped to pieces at uni if i don't check edge cases 2011-04-03T01:49:39 :P 2011-04-03T01:53:18 well, then I guess this is what I'm confused about. If you wanted to give an upper bound on what score could have been achieved by keeping the enemy alive then that would be "turns remaining * spawn rate + current food"... however this seems like such a large bonus that it would encourage stalemates 2011-04-03T01:53:52 so your argument was that it would not be such a large bonus? 2011-04-03T01:55:36 i was just trying to make sure people didn't get to some weird edge case and want to not fight just because the game was about to end 2011-04-03T01:55:44 like what we had with planet wars 2011-04-03T01:55:59 i won quite a few games in the final contest just because i knew to stop taking neutrals from the max turns 2011-04-03T01:56:12 like it actually happened more than once for me 2011-04-03T01:56:22 which is quite a few for such a seemingly small edge case 2011-04-03T01:57:44 ideally i would like max turns to be there just because it needs to be, and not really change the strategy people use as they approach it 2011-04-03T01:58:20 I'm not talking about the edge case anymore... I thought we agreed that would be handled in the obvious way :( 2011-04-03T01:58:24 alas that isn't the case with scoring, as if there's lots of players left, people are going to start killing like mad 2011-04-03T01:58:47 oh, hmmm sorry :P hang on 2011-04-03T01:59:18 i had no argument on that, but i do think it was too large before 2011-04-03T01:59:26 i don't mind not including current food 2011-04-03T01:59:54 you also need to remember that this player doesn't really know if they're competing with the last remaining bot or another bot for positions 2011-04-03T02:00:14 so i think they will go for the food bonus if they have a half decent chance 2011-04-03T02:00:30 hmmm... are you sure? 2011-04-03T02:01:21 i don't think the last two remaining bots will necessarily be the top 2 ranking bots 2011-04-03T02:01:34 the scoring is based around how much you have participated in the actual battles 2011-04-03T02:01:53 well, the food bonus upsets that :p 2011-04-03T02:01:56 i think a half decent "try to survive bot" might make it to the end of quite a few games 2011-04-03T02:02:06 yes, but that's just another part of the game 2011-04-03T02:02:21 I thought you didn't want that 2011-04-03T02:02:32 as that leads to stalemates 2011-04-03T02:02:32 maybe it should be you get a point for every ant you spawn? 2011-04-03T02:02:47 or you get a point for every ant every turn 2011-04-03T02:02:59 i don't want that 2011-04-03T02:03:03 I thought about that 2011-04-03T02:03:10 my point is that bot wouldn't win, so it's not a good strategy 2011-04-03T02:03:23 but if a person makes such a bot, then it may make it to such a situation without winning 2011-04-03T02:03:36 and with thousands of bots, there's bound to be plenty of variation 2011-04-03T02:04:03 but food bonus makes a "try to survive bot" strategy valid? 2011-04-03T02:04:08 does it not? 2011-04-03T02:04:15 it makes it more valid with the scoring system 2011-04-03T02:04:24 where's stocha when i need him 2011-04-03T02:04:33 i wouldn't mind having him try out those two other scoring methods above 2011-04-03T02:04:59 two other? how many scoring systems are there? 2011-04-03T02:05:05 there is the current one 2011-04-03T02:05:23 current is ""turns remaining * spawn rate"? 2011-04-03T02:05:23 then there is 1. you get an point for every ant spawned, 2. you get a point for every live ant every turn 2011-04-03T02:05:32 ah ok 2011-04-03T02:05:56 i kind of like 2 2011-04-03T02:06:08 although it may discourage fighting 2011-04-03T02:06:33 1. might be better, as you fight for territory to spawn more ants 2011-04-03T02:06:37 2 exacerbates the problem, you get a HUGE bonus for being the last alive 2011-04-03T02:07:08 just since it hasn't even been mentioned lately, I still prefer inverse order of elimination. ;) 2011-04-03T02:07:09 yeah, i don't like 2 30 seconds later 2011-04-03T02:07:26 janzert: that is simple 2011-04-03T02:07:47 actually, that way scoring doesn't determine what works best out of those various strategies 2011-04-03T02:07:57 (1) is similar to food bonus 2011-04-03T02:07:57 what gets you surviving the longest does 2011-04-03T02:08:09 whether that's camping, expanding, being catious, being fierce etc. 2011-04-03T02:08:15 * janzert nods 2011-04-03T02:08:25 * janzert goes away again 2011-04-03T02:08:35 sigh: what do you think about that? 2011-04-03T02:08:40 it's dead simple to explain too 2011-04-03T02:09:25 I agree with with what's been said... it is very appealing and the danger is that it leads to a stalemate 2011-04-03T02:09:45 yeah i was just running back to my computer to say that :) 2011-04-03T02:10:06 but if you don't fight and others expand, then you will be too small to compete 2011-04-03T02:10:13 with the fog of war it's hard to tell 2011-04-03T02:10:36 last alive wins, with the rest sorted by score might be a better compromise 2011-04-03T02:10:49 i don't think stale mates will be too much of an issue except with small openings where neither play can see an opportunity to do anything without getting pwned 2011-04-03T02:12:35 i think I prefer scoring the losers based on score so that you can't just get 2nd place by camping 2011-04-03T02:15:23 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-04-03T02:35:26 sigh: so no food bonus, just if there is one remaining participating player they automatically win? 2011-04-03T02:35:30 that's a bit weird 2011-04-03T02:35:41 the rest of the results could be wildly different 2011-04-03T02:36:40 not really, it's not that weird... it's the limit at the length of the game approaches infinity :P 2011-04-03T02:36:58 i just found another reason for not liking the food bonus 2011-04-03T02:37:13 what's that? 2011-04-03T02:37:13 depending on the map, it's going to have a different level of importance 2011-04-03T02:37:18 of course 2011-04-03T02:37:50 and depends on length of the game 2011-04-03T02:38:10 although we really need to get some decisions made and get the tournament going 2011-04-03T02:38:31 :P 2011-04-03T02:38:52 i'm going to start a wiki tutorial after 2011-04-03T02:39:16 atm i'm thinking of sections for path finding, food collection and game trees for battle 2011-04-03T02:39:42 this is the "advanced guide" not tutorial sorry 2011-04-03T02:40:10 i will do a general basic tutorial, although hopefully people will do a basic one specific to their starter packages 2011-04-03T02:40:21 cool 2011-04-03T02:40:41 feel free to give me ideas :P 2011-04-03T02:41:05 I'll be moving to sydney in like 2 weeks so I won't be looking at this stuff much longer :P 2011-04-03T02:41:33 yep, fair enough 2011-04-03T02:41:41 i'm starting to get busy too 2011-04-03T02:41:49 so i'm wanting to get my "committments" done 2011-04-03T02:41:57 not sure what my internet will be like when I get there either :P 2011-04-03T02:42:01 yeah 2011-04-03T02:43:29 *** phire has joined #aichallenge 2011-04-03T02:48:58 sigh: couldn't get internet worse than mine 2011-04-03T02:49:13 i'm on 1500/256 with a 20gb/month limit D: 2011-04-03T02:50:06 antimatroid, switch to a better isp 2011-04-03T02:50:15 you can actually get 1TB/month plans 2011-04-03T02:50:31 i live in tasmania :P 2011-04-03T02:50:45 and i live with my father, who "likes the convenience of one phone/internet bill" 2011-04-03T02:51:46 i shouldn't be that far away from being able to get adsl2 at least 2011-04-03T02:53:40 well thats one good thing about new zealand, its pretty hard to not get adsl2 2011-04-03T02:54:24 its pretty close to nationwide 2011-04-03T03:08:38 antimatroid: then you should offer to pay for internet and upgrade to something decent :P 2011-04-03T03:09:07 phire: i do like nz 2011-04-03T03:09:18 it's my backup plan if australia keeps going down the same track 2011-04-03T03:10:02 sigh: you say that like i ever have any money :P 2011-04-03T03:10:17 ah well, then you can't really complain :) 2011-04-03T03:18:40 *** Accoun has quit IRC () 2011-04-03T03:23:06 *** Accoun has joined #aichallenge 2011-04-03T03:25:27 *** amstan_ has quit IRC (Remote host closed the connection) 2011-04-03T03:57:24 *** Frontier has joined #aichallenge 2011-04-03T03:59:42 *** delt0r_ has joined #aichallenge 2011-04-03T04:01:23 *** delt0r___ has quit IRC (Ping timeout: 255 seconds) 2011-04-03T04:08:31 *** antimatroid has quit IRC (Ping timeout: 248 seconds) 2011-04-03T04:42:03 *** Frontier has quit IRC (Quit: Leaving.) 2011-04-03T04:44:31 *** Frontier has joined #aichallenge 2011-04-03T04:47:39 someone proposed to have a list of replays available on the net in form of a json array. it would be generated by a parametrized php. from the last contest we can see that there is interest in downloading all replays that match certain criteria, even if it is only your own games. 2011-04-03T04:48:21 sounds like a good idea 2011-04-03T04:48:25 Maybe it has been discussed here already. 2011-04-03T04:50:54 I would be interested in the list of games in which a certain bot that is stronger than mine lost, so I can try to figure out a good strategy against it. And (I posted that on gihub already) a 'continue from' parameter, so you never load the complete list just to get a refresh on the last couple of new games. 2011-04-03T04:51:38 A list of all games played by a Ruby bot or a Python bot seems rather pointless to me though. 2011-04-03T04:53:01 but a list of all games played on certain map types would be. 2011-04-03T04:53:51 you mean a complete list off all games played on symmetrical maps for example? 2011-04-03T04:55:04 more percisce, like all maps that have narrow gates between player locations 2011-04-03T04:56:03 phire: You are asking for quite some backend support here :) 2011-04-03T04:56:53 well, I'm pointing out that a large backend might be needed to provide useful information 2011-04-03T05:00:20 I think it is a very special case you think about. This is more something the players should do themselves then. But some filtering on the server should be in place anayway to avoid high loads. 2011-04-03T05:01:35 One step is to only provide links in the list of games. Links to the map and the replay as separate entities. This way you can skip replays where the map doesn't match your own criteria 2011-04-03T05:01:52 yes 2011-04-03T05:10:10 Ok, I'll now implement the fullscreen mode in the visualizer correctly (Java standalone version). Unfortunately to get rid of the window decorations you have to let Java free the associated system resources in a separate thread first. That is a bit annoying when you need this to happen instantly. So the best way seems to be to create a new window and move the contents from the normal window into the fullscreen window. 2011-04-03T05:14:27 *** antimatroid has joined #aichallenge 2011-04-03T05:19:11 .. what's better to use, a set of a map? 2011-04-03T05:19:34 depends 2011-04-03T05:20:03 i want to be able to check whether something is in there, iterate through it and delete stuff from it 2011-04-03T05:21:24 that's a set, isn't it? 2011-04-03T05:23:26 yes 2011-04-03T05:23:28 Sets are guaranteed to perform operations of insertion, deletion, and testing whether an element is in it, in logarithmic time - O(log n). 2011-04-03T05:23:32 but i can do it with a map too yes? 2011-04-03T05:24:34 antimatroid: of course, but why would you use a map unless you actually had values associated with each item? 2011-04-03T05:24:37 yes 2011-04-03T05:25:49 sigh: if they were faster/more efficient 2011-04-03T05:25:54 hence my question :P 2011-04-03T05:25:57 i'll go with sets 2011-04-03T05:26:35 i'm scrapping the "grid" representation for my actual bot 2011-04-03T05:29:19 *** p4p4p5 has joined #aichallenge 2011-04-03T05:35:55 what are simple straight up changes people can change other than avoiding collisions? 2011-04-03T05:45:05 find food? 2011-04-03T05:45:31 and/or enemies 2011-04-03T05:46:05 explore/survey territory 2011-04-03T05:46:21 better pathfinding 2011-04-03T05:59:40 pathfinding is my starting point for the proper guide 2011-04-03T06:00:21 http://pastebin.com/nefyN5i6 2011-04-03T06:00:35 that's what i currently have for a "tutorial.txt", i may throw some of that out though 2011-04-03T06:02:19 shouldn't setup stuff be separated from the tutorial? 2011-04-03T06:02:59 that's a file i'm planning to just include, i could split it into two i guess 2011-04-03T06:03:04 i didn't really want too many .txt files 2011-04-03T06:03:49 they serve quite different purposes 2011-04-03T06:05:00 i don't know, anything tutorialish in that is really for people to understand the source more than anything 2011-04-03T06:05:37 not sure what you mean 2011-04-03T06:06:09 have two files, one that answers "how do I run the bot", and another for "how do I improve the bot" 2011-04-03T06:07:00 getting_started.txt and getting_better.txt? :P 2011-04-03T06:08:22 "how do I run the bot" should definely be the readme 2011-04-03T06:08:26 *README 2011-04-03T06:08:29 yeah true 2011-04-03T06:08:34 i'll put that stuff in there 2011-04-03T06:08:42 anyway gotta take my brother to the airport, back after 2011-04-03T06:08:48 cya 2011-04-03T06:16:27 sigh: I'll implement the json replay today or tomorrow, but remain compatible with what we currently have. If someone hates it, which i doubt, I don't have to rewrite the parser. 2011-04-03T06:16:38 :) 2011-04-03T06:18:29 Frontier: also if you haven't already, please look at how visualize_locally.py works in the replay branch and tell me if you think that it is fine 2011-04-03T06:19:27 and that offline.html temaplte is alright 2011-04-03T06:19:41